Search results for: dimensionality

Authors: Dona Shaini Abhilasha Nanayakkara, Kurugamage Jude Pravinda Gregory Perera

Abstract:

E-commerce platforms have revolutionized the shopping experience, offering convenient ways for consumers to make purchases. To improve interactions with customers and optimize marketing strategies, it is essential for businesses to understand user behavior, preferences, and needs on these platforms. This paper focuses on recommending businesses to customize interactions with users based on their behavioral patterns, leveraging data-driven analysis and machine learning techniques. Businesses can improve engagement and boost the adoption of e-commerce platforms by aligning behavioral patterns with user goals of usability and satisfaction. We propose TheAnalyzer, a clustering-based system designed to enhance business productivity by analyzing user-profiles and improving human-computer interaction. The Analyzer seamlessly integrates with business applications, collecting relevant data points based on users' natural interactions without additional burdens such as questionnaires or surveys. It defines five key user analytics as features for its dataset, which are easily captured through users' interactions with e-commerce platforms. This research presents a study demonstrating the successful distinction of users into specific groups based on the five key analytics considered by TheAnalyzer. With the assistance of domain experts, customized business rules can be attached to each group, enabling The Analyzer to influence business applications and provide an enhanced personalized user experience. The outcomes are evaluated quantitatively and qualitatively, demonstrating that utilizing TheAnalyzer’s capabilities can optimize business outcomes, enhance customer satisfaction, and drive sustainable growth. The findings of this research contribute to the advancement of personalized interactions in e-commerce platforms. By leveraging user behavioral patterns and analyzing both new and existing users, businesses can effectively tailor their interactions to improve customer satisfaction, loyalty and ultimately drive sales.

Keywords: data clustering, data standardization, dimensionality reduction, human computer interaction, user profiling

Procedia PDF Downloads 34

Authors: O. Ekrami, P. Claes, S. Van Dongen

Abstract:

Fluctuating asymmetry (FA) has been studied for many years as an indicator of developmental stability or ‘genetic quality’ based on the assumption that perfect symmetry is ideally the expected outcome for a bilateral organism. Further studies have also investigated the possible link between FA and attractiveness or levels of masculinity or femininity. These hypotheses have been mostly examined using 2D images, and the structure of interest is usually presented using a limited number of landmarks. Such methods have the downside of simplifying and reducing the dimensionality of the structure, which will in return increase the error of the analysis. In an attempt to reach more conclusive and accurate results, in this study we have used high-resolution 3D scans of human faces and have developed an algorithm to measure and localize FA, taking a spatially-dense approach. A symmetric spatially dense anthropometric mask with paired vertices is non-rigidly mapped on target faces using an Iterative Closest Point (ICP) registration algorithm. A set of 19 manually indicated landmarks were used to examine the precision of our mapping step. The protocol’s accuracy in measurement and localizing FA is assessed using simulated faces with known amounts of asymmetry added to them. The results of validation of our approach show that the algorithm is perfectly capable of locating and measuring FA in 3D simulated faces. With the use of such algorithm, the additional captured information on asymmetry can be used to improve the studies of FA as an indicator of fitness or attractiveness. This algorithm can especially be of great benefit in studies of high number of subjects due to its automated and time-efficient nature. Additionally, taking a spatially dense approach provides us with information about the locality of FA, which is impossible to obtain using conventional methods. It also enables us to analyze the asymmetry of a morphological structures in a multivariate manner; This can be achieved by using methods such as Principal Components Analysis (PCA) or Factor Analysis, which can be a step towards understanding the underlying processes of asymmetry. This method can also be used in combination with genome wide association studies to help unravel the genetic bases of FA. To conclude, we introduced an algorithm to study and analyze asymmetry in human faces, with the possibility of extending the application to other morphological structures, in an automated, accurate and multi-variate framework.

Keywords: developmental stability, fluctuating asymmetry, morphometrics, 3D image processing

Procedia PDF Downloads 110

Authors: Nassima Noufail, Sara Bouhali

Abstract:

In this work, we develop a semi-supervised solution for the purpose of action detection in videos and propose an efficient algorithm for video segmentation. The approach is divided into video segmentation, feature extraction, and classification. In the first part, a video is segmented into clips, and we used the K-means algorithm for this segmentation; our goal is to find groups based on similarity in the video. The application of k-means clustering into all the frames is time-consuming; therefore, we started by the identification of transition frames where the scene in the video changes significantly, and then we applied K-means clustering into these transition frames. We used two image filters, the gaussian filter and the Laplacian of Gaussian. Each filter extracts a set of features from the frames. The Gaussian filter blurs the image and omits the higher frequencies, and the Laplacian of gaussian detects regions of rapid intensity changes; we then used this vector of filter responses as an input to our k-means algorithm. The output is a set of cluster centers. Each video frame pixel is then mapped to the nearest cluster center and painted with a corresponding color to form a visual map. The resulting visual map had similar pixels grouped. We then computed a cluster score indicating how clusters are near each other and plotted a signal representing frame number vs. clustering score. Our hypothesis was that the evolution of the signal would not change if semantically related events were happening in the scene. We marked the breakpoints at which the root mean square level of the signal changes significantly, and each breakpoint is an indication of the beginning of a new video segment. In the second part, for each segment from part 1, we randomly selected a 16-frame clip, then we extracted spatiotemporal features using convolutional 3D network C3D for every 16 frames using a pre-trained model. The C3D final output is a 512-feature vector dimension; hence we used principal component analysis (PCA) for dimensionality reduction. The final part is the classification. The C3D feature vectors are used as input to a multi-class linear support vector machine (SVM) for the training model, and we used a multi-classifier to detect the action. We evaluated our experiment on the UCF101 dataset, which consists of 101 human action categories, and we achieved an accuracy that outperforms the state of art by 1.2%.

Keywords: video segmentation, action detection, classification, Kmeans, C3D

Procedia PDF Downloads 43

Authors: Nikolai N. Bogolubov, Andrey V. Soldatov

Abstract:

Terahertz radiation occupies a range of frequencies somewhere from 100 GHz to approximately 10 THz, just between microwaves and infrared waves. This range of frequencies holds promise for many useful applications in experimental applied physics and technology. At the same time, reliable, simple techniques for generation, amplification, and modulation of electromagnetic radiation in this range are far from been developed enough to meet the requirements of its practical usage, especially in comparison to the level of technological abilities already achieved for other domains of the electromagnetic spectrum. This situation of relative underdevelopment of this potentially very important range of electromagnetic spectrum is known under the name of the 'terahertz gap.' Among other things, technological progress in the terahertz area has been impeded by the lack of compact, low energy consumption, easily controlled and continuously radiating terahertz radiation sources. Therefore, development of new techniques serving this purpose as well as various devices based on them is of obvious necessity. No doubt, it would be highly advantageous to employ the simplest of suitable physical systems as major critical components in these techniques and devices. The purpose of the present research was to show by means of conventional methods of non-equilibrium statistical mechanics and the theory of open quantum systems, that a thoroughly studied two-level quantum system, also known as an one-electron two-level 'atom', being driven by external classical monochromatic high-frequency (e.g. laser) field, can radiate continuously at much lower (e.g. terahertz) frequency in the fluorescent regime if the transition dipole moment operator of this 'atom' possesses permanent non-equal diagonal matrix elements. This assumption contradicts conventional assumption routinely made in quantum optics that only the non-diagonal matrix elements persist. The conventional assumption is pertinent to natural atoms and molecules and stems from the property of spatial inversion symmetry of their eigenstates. At the same time, such an assumption is justified no more in regard to artificially manufactured quantum systems of reduced dimensionality, such as, for example, quantum dots, which are often nicknamed 'artificial atoms' due to striking similarity of their optical properties to those ones of the real atoms. Possible ways to experimental observation and practical implementation of the predicted effect are discussed too.

Keywords: terahertz gap, two-level atom, resonant fluorescence, quantum dot, resonant fluorescence, two-level atom

Procedia PDF Downloads 236

Authors: Christian Neunzig, Simon Fahle, Jürgen Schulz, Matthias Möller, Bernd Kuhlenkötter

Abstract:

Digitalisation in production technology is a driver for the application of machine learning methods. Through the application of predictive quality, the great potential for saving necessary quality control can be exploited through the data-based prediction of product quality and states. However, the serial use of machine learning applications is often prevented by various problems. Fluctuations occur in real production data sets, which are reflected in trends and systematic shifts over time. To counteract these problems, data preprocessing includes rule-based data cleaning, the application of dimensionality reduction techniques, and the identification of comparable data subsets to extract stable features. Successful process control of the target variables aims to centre the measured values around a mean and minimise variance. Competitive leaders claim to have mastered their processes. As a result, much of the real data has a relatively low variance. For the training of prediction models, the highest possible generalisability is required, which is at least made more difficult by this data availability. The implementation of a machine learning application can be interpreted as a production process. The CRoss Industry Standard Process for Data Mining (CRISP-DM) is a process model with six phases that describes the life cycle of data science. As in any process, the costs to eliminate errors increase significantly with each advancing process phase. For the quality prediction of hydraulic test steps of directional control valves, the question arises in the initial phase whether a regression or a classification is more suitable. In the context of this work, the initial phase of the CRISP-DM, the business understanding, is critically compared for the use case at Bosch Rexroth with regard to regression and classification. The use of cross-process production data along the value chain of hydraulic valves is a promising approach to predict the quality characteristics of workpieces. Suitable methods for leakage volume flow regression and classification for inspection decision are applied. Impressively, classification is clearly superior to regression and achieves promising accuracies.

Keywords: classification, CRISP-DM, machine learning, predictive quality, regression

Procedia PDF Downloads 108

Authors: Maxence Queyrel, Edi Prifti, Alexandre Templier, Jean-Daniel Zucker

Abstract:

Analysis of the human microbiome using metagenomic sequencing data has demonstrated high ability in discriminating various human diseases. Raw metagenomic sequencing data require multiple complex and computationally heavy bioinformatics steps prior to data analysis. Such data contain millions of short sequences read from the fragmented DNA sequences and stored as fastq files. Conventional processing pipelines consist in multiple steps including quality control, filtering, alignment of sequences against genomic catalogs (genes, species, taxonomic levels, functional pathways, etc.). These pipelines are complex to use, time consuming and rely on a large number of parameters that often provide variability and impact the estimation of the microbiome elements. Training Deep Neural Networks directly from raw sequencing data is a promising approach to bypass some of the challenges associated with mainstream bioinformatics pipelines. Most of these methods use the concept of word and sentence embeddings that create a meaningful and numerical representation of DNA sequences, while extracting features and reducing the dimensionality of the data. In this paper we present an end-to-end approach that classifies patients into disease groups directly from raw metagenomic reads: metagenome2vec. This approach is composed of four steps (i) generating a vocabulary of k-mers and learning their numerical embeddings; (ii) learning DNA sequence (read) embeddings; (iii) identifying the genome from which the sequence is most likely to come and (iv) training a multiple instance learning classifier which predicts the phenotype based on the vector representation of the raw data. An attention mechanism is applied in the network so that the model can be interpreted, assigning a weight to the influence of the prediction for each genome. Using two public real-life data-sets as well a simulated one, we demonstrated that this original approach reaches high performance, comparable with the state-of-the-art methods applied directly on processed data though mainstream bioinformatics workflows. These results are encouraging for this proof of concept work. We believe that with further dedication, the DNN models have the potential to surpass mainstream bioinformatics workflows in disease classification tasks.

Keywords: deep learning, disease prediction, end-to-end machine learning, metagenomics, multiple instance learning, precision medicine

Procedia PDF Downloads 91

Authors: Manuel Salado, Mikel Rincón, Arkaitz Fidalgo, Roberto Fernandez, Senentxu Lanceros-Méndez

Abstract:

Lithium-ion batteries (LIBs) are a promising technology for energy storage, but they suffer from safety concerns due to the use of flammable organic solvents in their liquid electrolytes. Solid-state electrolytes (SSEs) offer a potential solution to this problem, but they have their own limitations, such as poor ionic conductivity and high interfacial resistance. The aim of this research was to develop a new type of SSE based on metal-organic frameworks (MOFs) and ionic liquids (ILs). MOFs are porous materials with high surface area and tunable electronic properties, making them ideal for use in SSEs. ILs are liquid electrolytes that are non-flammable and have high ionic conductivity. A series of MOFs were synthesized, and their electrochemical properties were evaluated. The MOFs were then infiltrated with ILs to form a quasi-solid gel and solid xerogel SSEs. The ionic conductivity, interfacial resistance, and electrochemical performance of the SSEs were characterized. The results showed that the MOF-IL SSEs had significantly higher ionic conductivity and lower interfacial resistance than conventional SSEs. The SSEs also exhibited excellent electrochemical performance, with high discharge capacity and long cycle life. The development of MOF-IL SSEs represents a significant advance in the field of solid-state electrolytes. The high ionic conductivity and low interfacial resistance of the SSEs make them promising candidates for use in next-generation LIBs. The data for this research was collected using a variety of methods, including X-ray diffraction, scanning electron microscopy, and electrochemical impedance spectroscopy. The data was analyzed using a variety of statistical and computational methods, including principal component analysis, density functional theory, and molecular dynamics simulations. The main question addressed by this research was whether MOF-IL SSEs could be developed that have high ionic conductivity, low interfacial resistance, and excellent electrochemical performance. The results of this research demonstrate that MOF-IL SSEs are a promising new type of solid-state electrolyte for use in LIBs. The SSEs have high ionic conductivity, low interfacial resistance, and excellent electrochemical performance. These properties make them promising candidates for use in next-generation LIBs that are safer and have higher energy densities.

Keywords: energy storage, solid-electrolyte, ionic liquid, metal-organic-framework, electrochemistry, organic inorganic plastic crystal

Procedia PDF Downloads 35

Authors: Chen Guo, Heng Tang, Ben Niu

Abstract:

Clustering splits objects into different groups based on similarity, making the objects have higher similarity in the same group and lower similarity in different groups. Thus, clustering can be treated as an optimization problem to maximize the intra-cluster similarity or inter-cluster dissimilarity. In real-world applications, the datasets often have some complex characteristics: sparse, overlap, high dimensionality, etc. When facing these datasets, simultaneously optimizing two or more objectives can obtain better clustering results than optimizing one objective. However, except for the objectives weighting methods, traditional clustering approaches have difficulty in solving multi-objective data clustering problems. Due to this, evolutionary multi-objective optimization algorithms are investigated by researchers to optimize multiple clustering objectives. In this paper, the Data Clustering algorithm based on Multi-objective Periodic Bacterial Foraging Optimization with two Learning Archives (DC-MPBFOLA) is proposed. Specifically, first, to reduce the high computing complexity of the original BFO, periodic BFO is employed as the basic algorithmic framework. Then transfer the periodic BFO into a multi-objective type. Second, two learning strategies are proposed based on the two learning archives to guide the bacterial swarm to move in a better direction. On the one hand, the global best is selected from the global learning archive according to the convergence index and diversity index. On the other hand, the personal best is selected from the personal learning archive according to the sum of weighted objectives. According to the aforementioned learning strategies, a chemotaxis operation is designed. Third, an elite learning strategy is designed to provide fresh power to the objects in two learning archives. When the objects in these two archives do not change for two consecutive times, randomly initializing one dimension of objects can prevent the proposed algorithm from falling into local optima. Fourth, to validate the performance of the proposed algorithm, DC-MPBFOLA is compared with four state-of-art evolutionary multi-objective optimization algorithms and one classical clustering algorithm on evaluation indexes of datasets. To further verify the effectiveness and feasibility of designed strategies in DC-MPBFOLA, variants of DC-MPBFOLA are also proposed. Experimental results demonstrate that DC-MPBFOLA outperforms its competitors regarding all evaluation indexes and clustering partitions. These results also indicate that the designed strategies positively influence the performance improvement of the original BFO.

Keywords: data clustering, multi-objective optimization, bacterial foraging optimization, learning archives

Procedia PDF Downloads 103

Authors: Zunaira Zaman, David Bohannon

Abstract:

This study explores the application of machine learning (ML) for detecting security vulnerabilities in source code. The research aims to assist organizations with large application portfolios and limited security testing capabilities in prioritizing security activities. ML-based approaches offer benefits such as increased confidence scores, false positives and negatives tuning, and automated feedback. The initial approach using natural language processing techniques to extract features achieved 86% accuracy during the training phase but suffered from overfitting and performed poorly on unseen datasets during testing. To address these issues, the study proposes using the abstract syntax tree (AST) for Java and C++ codebases to capture code semantics and structure and generate path-context representations for each function. The Code2Vec model architecture is used to learn distributed representations of source code snippets for training a machine-learning classifier for vulnerability prediction. The study evaluates the performance of the proposed methodology using two datasets and compares the results with existing approaches. The Devign dataset yielded 60% accuracy in predicting vulnerable code snippets and helped resist overfitting, while the Juliet Test Suite predicted specific vulnerabilities such as OS-Command Injection, Cryptographic, and Cross-Site Scripting vulnerabilities. The Code2Vec model achieved 75% accuracy and a 98% recall rate in predicting OS-Command Injection vulnerabilities. The study concludes that even partial AST representations of source code can be useful for vulnerability prediction. The approach has the potential for automated intelligent analysis of source code, including vulnerability prediction on unseen source code. State-of-the-art models using natural language processing techniques and CNN models with ensemble modelling techniques did not generalize well on unseen data and faced overfitting issues. However, predicting vulnerabilities in source code using machine learning poses challenges such as high dimensionality and complexity of source code, imbalanced datasets, and identifying specific types of vulnerabilities. Future work will address these challenges and expand the scope of the research.

Keywords: code embeddings, neural networks, natural language processing, OS command injection, software security, code properties

Procedia PDF Downloads 70

Authors: Mostafa Shahriari, Theophile Chaumont-Frelet, David Pardo

Abstract:

Resistivity measurements are used to characterize the Earth’s subsurface. They are categorized into two different groups: (a) those acquired on the Earth’s surface, for instance, controlled source electromagnetic (CSEM) and Magnetotellurics (MT), and (b) those recorded with borehole logging instruments such as Logging-While-Drilling (LWD) devices. LWD instruments are mostly used for geo-steering purposes, i.e., to adjust dip and azimuthal angles of a well trajectory to drill along a particular geological target. Modern LWD tools measure all nine components of the magnetic field corresponding to three orthogonal transmitter and receiver orientations. In order to map the Earth’s subsurface and perform geo-steering, we invert measurements using a gradient-based method that utilizes the derivatives of the recorded measurements with respect to the inversion variables. For resistivity measurements, these inversion variables are usually the constant resistivity value of each layer and the bed boundary positions. It is well-known how to compute derivatives with respect to the constant resistivity value of each layer using semi-analytic or numerical methods. However, similar formulas for computing the derivatives with respect to bed boundary positions are unavailable. The main contribution of this work is to provide an adjoint-based formulation for computing derivatives with respect to the bed boundary positions. The key idea to obtain the aforementioned adjoint state formulations for the derivatives is to separate the tangential and normal components of the field and treat them differently. This formulation allows us to compute the derivatives faster and more accurately than with traditional finite differences approximations. In the presentation, we shall first derive a formula for computing the derivatives with respect to the bed boundary positions for the potential equation. Then, we shall extend our formulation to 3D Maxwell’s equations. Finally, by considering a 1D domain and reducing the dimensionality of the problem, which is a common practice in the inversion of resistivity measurements, we shall derive a formulation to compute the derivatives of the measurements with respect to the bed boundary positions using a 1.5D variational formulation. Then, we shall illustrate the accuracy and convergence properties of our formulations by comparing numerical results with the analytical derivatives for the potential equation. For the 1.5D Maxwell’s system, we shall compare our numerical results based on the proposed adjoint-based formulation vs those obtained with a traditional finite difference approach. Numerical results shall show that our proposed adjoint-based technique produces enhanced accuracy solutions while its cost is negligible, as opposed to the finite difference approach that requires the solution of one additional problem per derivative.

Keywords: inverse problem, bed boundary positions, electromagnetism, potential equation

Procedia PDF Downloads 152

Authors: Chua Si Wen Alicia, Marcus Goh Tian Xi, Eunice Gan Ghee Wu, Helen Bound, Lee Liang Ying, Albert Lee

Abstract:

In a world full of rapid changes, people often need a set of skills to help them navigate an ever-changing workscape. These skills, often known as “future-oriented skills,” include learning to learn, critical thinking, understanding multiple perspectives, and knowledge creation. Future-oriented skills are typically assumed to be domain-general, applicable to multiple domains, and can be cultivated through a learning approach called Dialogical Inquiry. Dialogical Inquiry is known for its benefits of making sense of multiple perspectives, encouraging critical thinking, and developing learner’s capability to learn. However, it currently exists as a quantitative tool, which makes it hard to track and compare learning processes over time. With these concerns, the present research aimed to develop and validate a quantitative tool for the Map of Dialogical Inquiry, focusing Imagining aspect of learning. The Imagining aspect four dimensions: 1) speculative/ look for alternatives, 2) risk taking/ break rules, 3) create/ design, and 4) vision/ imagine. To do so, an exploratory literature review was conducted to better understand the dimensions of Imagining. This included deep-diving into the history of the creation of the Map of Dialogical Inquiry and a review on how “Imagining” has been conceptually defined in the field of social psychology, education, and beyond. Then, we synthesised and validated scales. These scales measured the dimension of Imagination and related concepts like creativity, divergent thinking regulatory focus, and instrumental risk. Thereafter, items were adapted from the aforementioned procured scales to form items that would contribute to the preliminary version of the Imagining Scale. For scale validation, 250 participants were recruited. A Confirmatory Factor Analysis (CFA) sought to establish dimensionality of the Imagining Scale with an iterative procedure in item removal. Reliability and validity of the scale’s dimensions were sought through measurements of Cronbach’s alpha, convergent validity, and discriminant validity. While CFA found that the distinction of Imagining’s four dimensions could not be validated, the scale was able to establish high reliability with a Cronbach alpha of .96. In addition, the convergent validity of the Imagining scale was established. A lack of strong discriminant validity may point to overlaps with other components of the Dialogical Map as a measure of learning. Thus, a holistic approach to forming the tool – encompassing all eight different components may be preferable.

Keywords: learning, education, imagining, pedagogy, dialogical teaching

Procedia PDF Downloads 60

Authors: Eunice Gan Ghee Wu, Marcus Goh Tian Xi, Alicia Chua Si Wen, Helen Bound, Lee Liang Ying, Albert Lee

Abstract:

The Map of Dialogical Inquiry provides a conceptual basis of learning processes. According to the Map, dialogical inquiry motivates complex thinking, dialogue, reflection, and learner agency. For instance, classrooms that incorporated dialogical inquiry enabled learners to construct more meaning in their learning, to engage in self-reflection, and to challenge their ideas with different perspectives. While the Map contributes to the psychology of learning, its qualitative approach makes it hard to track and compare learning processes over time for both teachers and learners. Qualitative approach typically relies on open-ended responses, which can be time-consuming and resource-intensive. With these concerns, the present research aimed to develop and validate a quantifiable measure for the Map. Specifically, the Map of Dialogical Inquiry reflects the eight different learning processes and perspectives employed during a learner’s experience. With a focus on interpersonal and emotional learning processes, the purpose of the present study is to construct and validate a scale to measure the “Relating” aspect of learning. According to the Map, the Relating aspect of learning contains four conceptual components: using intuition and empathy, seeking personal meaning, building relationships and meaning with others, and likes stories and metaphors. All components have been shown to benefit learning in past research. This research began with a literature review with the goal of identifying relevant scales in the literature. These scales were used as a basis for item development, guided by the four conceptual dimensions in the “Relating” aspect of learning, resulting in a pool of 47 preliminary items. Then, all items were administered to 200 American participants via an online survey along with other scales of learning. Dimensionality, reliability, and validity of the “Relating” scale was assessed. Data were submitted to a confirmatory factor analysis (CFA), revealing four distinct components and items. Items with lower factor loadings were removed in an iterative manner, resulting in 34 items in the final scale. CFA also revealed that the “Relating” scale was a four-factor model, following its four distinct components as described in the Map of Dialogical Inquiry. In sum, this research was able to develop a quantitative scale for the “Relating” aspect of the Map of Dialogical Inquiry. By representing learning as numbers, users, such as educators and learners, can better track, evaluate, and compare learning processes over time in an efficient manner. More broadly, this scale may also be used as a learning tool in lifelong learning.

Keywords: lifelong learning, scale development, dialogical inquiry, relating, social and emotional learning, socio-affective intuition, empathy, narrative identity, perspective taking, self-disclosure

Procedia PDF Downloads 104

Authors: Mohamed Alimoussa, Alice Porebski, Nicolas Vandenbroucke, Rachid Oulad Haj Thami, Sana El Fkihi

Abstract:

Color and texture are highly discriminant visual cues that provide an essential information in many types of images. Color texture representation and classification is therefore one of the most challenging problems in computer vision and image processing applications. Color textures can be represented in different color spaces by using multiple image descriptors which generate a high dimensional set of texture features. In order to reduce the dimensionality of the feature set, feature selection techniques can be used. The goal of feature selection is to find a relevant subset from an original feature space that can improve the accuracy and efficiency of a classification algorithm. Traditionally, feature selection is focused on removing irrelevant features, neglecting the possible redundancy between relevant ones. This is why some feature selection approaches prefer to use feature clustering analysis to aid and guide the search. These techniques can be divided into two categories. i) Feature clustering-based ranking algorithm uses feature clustering as an analysis that comes before feature ranking. Indeed, after dividing the feature set into groups, these approaches perform a feature ranking in order to select the most discriminant feature of each group. ii) Feature clustering-based subset search algorithms can use feature clustering following one of three strategies; as an initial step that comes before the search, binded and combined with the search or as the search alternative and replacement. In this paper, we propose a new feature clustering-based sequential selection approach for the purpose of color texture representation and classification. Our approach is a three step algorithm. First, irrelevant features are removed from the feature set thanks to a class-correlation measure. Then, introducing a new automatic feature clustering algorithm, the feature set is divided into several feature clusters. Finally, a sequential search algorithm, based on a filter model and a separability measure, builds a relevant and non redundant feature subset: at each step, a feature is selected and features of the same cluster are removed and thus not considered thereafter. This allows to significantly speed up the selection process since large number of redundant features are eliminated at each step. The proposed algorithm uses the clustering algorithm binded and combined with the search. Experiments using a combination of two well known texture descriptors, namely Haralick features extracted from Reduced Size Chromatic Co-occurence Matrices (RSCCMs) and features extracted from Local Binary patterns (LBP) image histograms, on five color texture data sets, Outex, NewBarktex, Parquet, Stex and USPtex demonstrate the efficiency of our method compared to seven of the state of the art methods in terms of accuracy and computation time.

Keywords: feature selection, color texture classification, feature clustering, color LBP, chromatic cooccurrence matrix

Procedia PDF Downloads 94

Authors: Abdul-Ganiyu Fuseini, Bernice Redley, Helen Rawson, Lenore Lay, Debra Kerr

Abstract:

Aim: The study aimed to develop and validate a culturally appropriate patient-reported outcome measure for measuring dignity for older adults during acute hospital admissions. Design: A three-phased mixed-method sequential exploratory design was used. Methods: Concept elicitation and generation of items for the scale was informed by older adults’ perspectives about dignity during acute hospitalization and a literature review. Content validity evaluation and pre-testing were undertaken using standard instrument development techniques. A cross-sectional survey design was conducted involving 270 hospitalized older adults for evaluation of construct and convergent validity, internal consistency reliability, and test–retest reliability of the scale. Analysis was performed using Statistical Package for the Social Sciences, version 25. Reporting of the study was guided by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist. Results: We established the 15-item Hospitalized Older Adults’ Dignity Scale that has a 5-factor structure: Shared Decision-Making (3 items); Healthcare Professional-Patient Communication (3 items); Patient Autonomy (4 items); Patient Privacy (2 items); and Respectful Care (3 items). Excellent content validity, adequate construct and convergent validity, acceptable internal consistency reliability, and good test-retest reliability were demonstrated. Conclusion: We established the Hospitalized Older Adults Dignity Scale as a valid and reliable scale to measure dignity for older adults during acute hospital admissions. Future studies using confirmatory factor analysis are needed to corroborate the dimensionality of the factor structure and external validity of the scale. Routine use of the scale may provide information that informs the development of strategies to improve dignity-related care in the future. Impact: The development and validation of the Hospitalized Older Adults Dignity Scale will provide healthcare professionals with a feasible and reliable scale for measuring older adults’ dignity during acute hospitalization. Routine use of the scale may enable the capturing and incorporation of older patients’ perspectives about their healthcare experience and provide information that informs the development of strategies to improve dignity-related care in the future.

Keywords: dignity, older adults, hospitalisation, scale, patients, dignified care, acute care

Procedia PDF Downloads 61

Authors: Amani Tahat, Serdal Kirmizialtin

Abstract:

Trajectory fragment methods such as Markov State Models (MSM), Milestoning (MS) and Transition Path sampling are the prime choice of extending the timescale of all atom Molecular Dynamics simulations. In these approaches, a set of structures that covers the accessible phase space has to be chosen a priori using cluster analysis. Structural clustering serves to partition the conformational state into natural subgroups based on their similarity, an essential statistical methodology that is used for analyzing numerous sets of empirical data produced by Molecular Dynamics (MD) simulations. Local transition kernel among these clusters later used to connect the metastable states using a Markovian kinetic model in MSM and a non-Markovian model in MS. The choice of clustering approach in constructing such kernel is crucial since the high dimensionality of the biomolecular structures might easily confuse the identification of clusters when using the traditional hierarchical clustering methodology. Of particular interest, in the case of MS where the milestones are very close to each other, accurate determination of the milestone identity of the trajectory becomes a challenging issue. Throughout this work we present two cluster analysis methods applied to the cis–trans isomerism of dinucleotide AA. The choice of nucleic acids to commonly used proteins to study the cluster analysis is two fold: i) the energy landscape is rugged; hence transitions are more complex, enabling a more realistic model to study conformational transitions, ii) Nucleic acids conformational space is high dimensional. A diverse set of internal coordinates is necessary to describe the metastable states in nucleic acids, posing a challenge in studying the conformational transitions. Herein, we need improved clustering methods that accurately identify the AA structure in its metastable states in a robust way for a wide range of confused data conditions. The single linkage approach of the hierarchical clustering available in GROMACS MD-package is the first clustering methodology applied to our data. Self Organizing Map (SOM) neural network, that also known as a Kohonen network, is the second data clustering methodology. The performance comparison of the neural network as well as hierarchical clustering method is studied by means of computing the mean first passage times for the cis-trans conformational rates. Our hope is that this study provides insight into the complexities and need in determining the appropriate clustering algorithm for kinetic analysis. Our results can improve the effectiveness of decisions based on clustering confused empirical data in studying conformational transitions in biomolecules.

Keywords: milestoning, self organizing map, single linkage, structure clustering

Procedia PDF Downloads 188

Authors: Jiangbo Li, Wenqian Huang

Abstract:

Early detection of fungal infection in citrus fruit is one of the major problems in the postharvest commercialization process. The automatic and nondestructive detection of infected fruits is still a challenge for the citrus industry. At present, the visual inspection of rotten citrus fruits is commonly performed by workers through the ultraviolet induction fluorescence technology or manual sorting in citrus packinghouses to remove fruit subject with fungal infection. However, the former entails a number of problems because exposing people to this kind of lighting is potentially hazardous to human health, and the latter is very inefficient. Orange is used as a research object. This study would focus on this problem and proposed an effective method based on Vis-NIR hyperspectral imaging in the wavelength range of 400-1000 nm with a spectroscopic resolution of 2.8 nm. In this work, three normalization approaches are applied prior to analysis to reduce the effect of sample curvature on spectral profiles, and it is found that mean normalization was the most effective pretreatment for decreasing spectral variability due to curvature. Then, principal component analysis (PCA) was applied to a dataset composing of average spectra from decayed and normal tissue to reduce the dimensionality of data and observe the ability of Vis-NIR hyper-spectra to discriminate data from two classes. In this case, it was observed that normal and decayed spectra were separable along the resultant first principal component (PC1) axis. Subsequently, five wavelengths (band) centered at 577, 702, 751, 808, and 923 nm were selected as the characteristic wavelengths by analyzing the loadings of PC1. A multispectral combination image was generated based on five selected characteristic wavelength images. Based on the obtained multispectral combination image, the intensity slicing pseudocolor image processing method is used to generate a 2-D visual classification image that would enhance the contrast between normal and decayed tissue. Finally, an image segmentation algorithm for detection of decayed fruit was developed based on the pseudocolor image coupled with a simple thresholding method. For the investigated 238 independent set samples including infected fruits infected by Penicillium digitatum and normal fruits, the total success rate is 100% and 97.5%, respectively, and, the proposed algorithm also used to identify the orange infected by penicillium italicum with a 100% identification accuracy, indicating that the proposed multispectral algorithm here is an effective method and it is potential to be applied in citrus industry.

Keywords: citrus fruit, early rotten, fungal infection, hyperspectral imaging

Procedia PDF Downloads 264

Authors: Jianan Sun, Ziwen Ye

Abstract:

Multidimensional computerized adaptive testing (MCAT) is a popular research topic in psychometrics. It is important for practitioners to clearly know the item-trait patterns of administered items when a test like MCAT is operated. Item-trait pattern recognition refers to detecting which latent traits in a psychological test are measured by each of the specified items. If the item-trait patterns of the replenished items in MCAT item pool are well detected, the interpretability of the items can be improved, which can further promote the abilities of the examinees who attending the MCAT to be accurately estimated. This research explores to solve the item-trait pattern recognition problem of the replenished items in MCAT item pool from the perspective of statistical variable selection. The popular multidimensional item response theory model, multidimensional two-parameter logistic model, is assumed to fit the response data of MCAT. The proposed method uses the least absolute shrinkage and selection operator (LASSO) to detect item-trait patterns of replenished items based on the essential information of item responses and ability estimates of examinees collected from a designed MCAT procedure. Several advantages of the proposed method are outlined. First, the proposed method does not strictly depend on the relative order between the replenished items and the selected operational items, so it allows the replenished items to be mixed into the operational items in reasonable order such as considering content constraints or other test requirements. Second, the LASSO used in this research improves the interpretability of the multidimensional replenished items in MCAT. Third, the proposed method can exert the advantage of shrinkage method idea for variable selection, so it can help to check item quality and key dimension features of replenished items and saves more costs of time and labors in response data collection than traditional factor analysis method. Moreover, the proposed method makes sure the dimensions of replenished items are recognized to be consistent with the dimensions of operational items in MCAT item pool. Simulation studies are conducted to investigate the performance of the proposed method under different conditions for varying dimensionality of item pool, latent trait correlation, item discrimination, test lengths and item selection criteria in MCAT. Results show that the proposed method can accurately detect the item-trait patterns of the replenished items in the two-dimensional and the three-dimensional item pool. Selecting enough operational items from the item pool consisting of high discriminating items by Bayesian A-optimality in MCAT can improve the recognition accuracy of item-trait patterns of replenished items for the proposed method. The pattern recognition accuracy for the conditions with correlated traits is better than those with independent traits especially for the item pool consisting of comparatively low discriminating items. To sum up, the proposed data-driven method based on the LASSO can accurately and efficiently detect the item-trait patterns of replenished items in MCAT.

Keywords: item-trait pattern recognition, least absolute shrinkage and selection operator, multidimensional computerized adaptive testing, variable selection

Procedia PDF Downloads 91

Authors: Ghita Benayad

Abstract:

Asset allocation, which affects the long-term profitability of portfolios by distributing assets to fulfill a range of investment objectives, is the cornerstone of investment management in the dynamic and complicated world of financial markets. This paper offers a technique for optimizing strategic asset allocation with the goal of improving portfolio performance by addressing the inherent complexity and uncertainty of the market through the use of Principal Component Analysis (PCA) in a multi-objective modeling framework. The study's first section starts with a critical evaluation of conventional asset allocation techniques, highlighting how poorly they are able to capture the intricate relationships between assets and the volatile nature of the market. In order to overcome these challenges, the project suggests a PCA-driven methodology that isolates important characteristics influencing asset returns by decreasing the dimensionality of the investment universe. This decrease provides a stronger basis for asset allocation decisions by facilitating a clearer understanding of market structures and behaviors. Using a multi-objective optimization model, the project builds on this foundation by taking into account a number of performance metrics at once, including risk minimization, return maximization, and the accomplishment of predetermined investment goals like regulatory compliance or sustainability standards. This model provides a more comprehensive understanding of investor preferences and portfolio performance in comparison to conventional single-objective optimization techniques. While applying the PCA-driven multi-objective optimization model to historical market data, aiming to construct portfolios better under different market situations. As compared to portfolios produced from conventional asset allocation methodologies, the results show that portfolios optimized using the proposed method display improved risk-adjusted returns, more resilience to market downturns, and better alignment with specified investment objectives. The study also looks at the implications of this PCA technique for portfolio management, including the prospect that it might give investors a more advanced framework for navigating financial markets. The findings suggest that by combining PCA with multi-objective optimization, investors may obtain a more strategic and informed asset allocation that is responsive to both market conditions and individual investment preferences. In conclusion, this capstone project improves the field of financial engineering by creating a sophisticated asset allocation optimization model that integrates PCA with multi-objective optimization. In addition to raising concerns about the condition of asset allocation today, the proposed method of portfolio management opens up new avenues for research and application in the area of investment techniques.

Keywords: asset allocation, portfolio optimization, principle component analysis, multi-objective modelling, financial market

Procedia PDF Downloads 14

Authors: Alexandru-Ion Marinescu

Abstract:

There exist a plethora of methods in the scientific literature which tackle the well-established task of credit score evaluation. In its most abstract form, a credit scoring algorithm takes as input several credit applicant properties, such as age, marital status, employment status, loan duration, etc. and must output a binary response variable (i.e. “GOOD” or “BAD”) stating whether the client is susceptible to payment return delays. Data imbalance is a common occurrence among financial institution databases, with the majority being classified as “GOOD” clients (clients that respect the loan return calendar) alongside a small percentage of “BAD” clients. But it is the “BAD” clients we are interested in since accurately predicting their behavior is crucial in preventing unwanted loss for loan providers. We add to this whole context the constraint that the algorithm must yield an actual, tractable mathematical formula, which is friendlier towards financial analysts. To this end, we have turned to genetic algorithms and genetic programming, aiming to evolve actual mathematical expressions using specially tailored mutation and crossover operators. As far as data representation is concerned, we employ a very flexible mechanism – LINQ expression trees, readily available in the C# programming language, enabling us to construct executable pieces of code at runtime. As the title implies, they model trees, with intermediate nodes being operators (addition, subtraction, multiplication, division) or mathematical functions (sin, cos, abs, round, etc.) and leaf nodes storing either constants or variables. There is a one-to-one correspondence between the client properties and the formula variables. The mutation and crossover operators work on a flattened version of the tree, obtained via a pre-order traversal. A consequence of our chosen technique is that we can identify and discard client properties which do not take part in the final score evaluation, effectively acting as a dimensionality reduction scheme. We compare ourselves with state of the art approaches, such as support vector machines, Bayesian networks, and extreme learning machines, to name a few. The data sets we benchmark against amount to a total of 8, of which we mention the well-known Australian credit and German credit data sets, and the performance indicators are the following: percentage correctly classified, area under curve, partial Gini index, H-measure, Brier score and Kolmogorov-Smirnov statistic, respectively. Finally, we obtain encouraging results, which, although placing us in the lower half of the hierarchy, drive us to further refine the algorithm.

Keywords: expression trees, financial credit scoring, genetic algorithm, genetic programming, symbolic evolution

Procedia PDF Downloads 88

Authors: Ryan A. Black, Stacey A. McCaffrey

Abstract:

Over the past few decades, great strides have been made towards improving the science in the measurement of psychological constructs. Item Response Theory (IRT) has been the foundation upon which statistical models have been derived to increase both precision and accuracy in psychological measurement. These models are now being used widely to develop and refine tests intended to measure an individual's level of academic achievement, aptitude, and intelligence. Recently, the field of clinical psychology has adopted IRT models to measure psychopathological phenomena such as depression, anxiety, and addiction. Because advances in IRT measurement models are being made so rapidly across various fields, it has become quite challenging for psychologists and other behavioral scientists to keep abreast of the most recent developments, much less learn how to employ and decide which models are the most appropriate to use in their line of work. In the same vein, IRT measurement models vary greatly in complexity in several interrelated ways including but not limited to the number of item-specific parameters estimated in a given model, the function which links the expected response and the predictor, response option formats, as well as dimensionality. As a result, inferior methods (a.k.a. Classical Test Theory methods) continue to be employed in efforts to measure psychological constructs, despite evidence showing that IRT methods yield more precise and accurate measurement. To increase the use of IRT methods, this study endeavors to provide a comprehensive overview of binary IRT models; that is, measurement models employed on test data consisting of binary response options (e.g., correct/incorrect, true/false, agree/disagree). Specifically, this study will cover the most basic binary IRT model, known as the 1-parameter logistic (1-PL) model dating back to over 50 years ago, up until the most recent complex, 4-parameter logistic (4-PL) model. Binary IRT models will be defined mathematically and the interpretation of each parameter will be provided. Next, all four binary IRT models will be employed on two sets of data: 1. Simulated data of N=500,000 subjects who responded to four dichotomous items and 2. A pilot analysis of real-world data collected from a sample of approximately 770 subjects who responded to four self-report dichotomous items pertaining to emotional consequences to alcohol use. Real-world data were based on responses collected on items administered to subjects as part of a scale-development study (NIDA Grant No. R44 DA023322). IRT analyses conducted on both the simulated data and analyses of real-world pilot will provide a clear demonstration of how to construct, evaluate, and compare binary IRT measurement models. All analyses will be performed using the new IRT procedure in SAS 9.4. SAS code to generate simulated data and analyses will be available upon request to allow for replication of results.

Keywords: instrument development, item response theory, latent trait theory, psychometrics

Procedia PDF Downloads 324

Authors: Aabha Chaubey, Bani Bhattacharya

Abstract:

Evaluation of the quality of teaching in engineering education in India needs to be conducted on a continuous basis to achieve the best teaching quality in technical education. Quality teaching is an influential factor in technical education which impacts largely on learning outcomes of the students. Present study is not exclusively theory-driven, but it draws on various specific concepts and constructs in the domain of technical education. These include teaching and learning in higher education, teacher effectiveness, and teacher evaluation and performance management in higher education. Student Evaluation of Education Quality (SEEQ) was proposed as one of the evaluation instruments of the quality teaching in engineering education. SEEQ is one of the popular and standard instrument widely utilized all over the world and bears the validity and reliability in educational world. The present study was designed to evaluate the teaching quality through SEEQ in the context of technical education in India, including its validity and reliability based on the collected data. The multiple dimensionality of SEEQ that is present in every teaching and learning process made it quite suitable to collect the feedback of students regarding the quality of instructions and instructor. The SEEQ comprises of 9 original constructs i.e.; learning value, teacher enthusiasm, organization, group interaction, and individual rapport, breadth of coverage, assessment, assignments and overall rating of particular course and instructor with total of 33 items. In the present study, a total of 350 samples comprising first year undergraduate students from Indian Institute of Technology, Kharagpur (IIT, Kharagpur, India) were included for the evaluation of the importance of SEEQ. They belonged to four different courses of different streams of engineering studies. The above studies depicted the validity and reliability of SEEQ was based upon the collected data. This further needs Confirmatory Factor Analysis (CFA) and Analysis of Moment structure (AMOS) for various scaled instrument like SEEQ Cronbach’s alpha which are associated with SPSS for the examination of the internal consistency. The evaluation of the effectiveness of SEEQ in CFA is implemented on the basis of fit indices such as CMIN/df, CFI, GFI, AGFI and RMSEA readings. The major findings of this study showed the fitness indices such as ChiSq = 993.664,df = 390,ChiSq/df = 2.548,GFI = 0.782,AGFI = 0.736,CFI = 0.848,RMSEA = 0.062,TLI = 0.945,RMR = 0.029,PCLOSE = 0.006. The final analysis of the fit indices presented positive construct validity and stability, on the other hand a higher reliability was also depicted which indicated towards internal consistency. Thus, the study suggests the effectivity of SEEQ as the indicator of the quality evaluation instrument in teaching-learning process in engineering education in India. Therefore, it is expected that with the continuation of this research in engineering education there remains a possibility towards the betterment of the quality of the technical education in India. It is also expected that this study will provide an empirical and theoretical logic towards locating a construct or factor related to teaching, which has the greatest impact on teaching and learning process in a particular course or stream in engineering education.

Keywords: confirmatory factor analysis, engineering education, SEEQ, teaching and learning process

Procedia PDF Downloads 389

Authors: Mostafa Shahriari, Sergio Rojas, David Pardo, Angel Rodriguez- Rozas, Shaaban A. Bakr, Victor M. Calo, Ignacio Muga

Abstract:

Logging-While Drilling (LWD) is a technique to record down-hole logging measurements while drilling the well. Nowadays, LWD devices (e.g., nuclear, sonic, resistivity) are mostly used commercially for geo-steering applications. Modern borehole resistivity tools are able to measure all components of the magnetic field by incorporating tilted coils. The depth of investigation of LWD tools is limited compared to the thickness of the geological layers. Thus, it is a common practice to approximate the Earth’s subsurface with a sequence of 1D models. For a 1D model, we can reduce the dimensionality of the problem using a Hankel transform. We can solve the resulting system of ordinary differential equations (ODEs) either (a) analytically, which results in a so-called semi-analytic method after performing a numerical inverse Hankel transform, or (b) numerically. Semi-analytic methods are used by the industry due to their high performance. However, they have major limitations, namely: -The analytical solution of the aforementioned system of ODEs exists only for piecewise constant resistivity distributions. For arbitrary resistivity distributions, the solution of the system of ODEs is unknown by today’s knowledge. -In geo-steering, we need to solve inverse problems with respect to the inversion variables (e.g., the constant resistivity value of each layer and bed boundary positions) using a gradient-based inversion method. Thus, we need to compute the corresponding derivatives. However, the analytical derivatives of cross-bedded formation and the analytical derivatives with respect to the bed boundary positions have not been published to the best of our knowledge. The main contribution of this work is to overcome the aforementioned limitations of semi-analytic methods by solving each 1D model (associated with each Hankel mode) using an efficient multi-scale finite element method. The main idea is to divide our computations into two parts: (a) offline computations, which are independent of the tool positions and we precompute only once and use them for all logging positions, and (b) online computations, which depend upon the logging position. With the above method, (a) we can consider arbitrary resistivity distributions along the 1D model, and (b) we can easily and rapidly compute the derivatives with respect to any inversion variable at a negligible additional cost by using an adjoint state formulation. Although the proposed method is slower than semi-analytic methods, its computational efficiency is still high. In the presentation, we shall derive the mathematical variational formulation, describe the proposed multi-scale finite element method, and verify the accuracy and efficiency of our method by performing a wide range of numerical experiments and comparing the numerical solutions to semi-analytic ones when the latest are available.

Keywords: logging-While-Drilling, resistivity measurements, multi-scale finite elements, Hankel transform

Procedia PDF Downloads 353

Authors: Angelo Corallo, Maria Elena Latino, Marta Menegoli

Abstract:

The link between man and food plays, in the social and economic system, a central role where cultural and multidisciplinary aspects intertwine: food is not only nutrition, but also communication, culture, politics, environment, science, ethics, fashion. This multi-dimensionality has many implications in the food economy. In recent years, the consumer became more conscious about his food choices, involving a consistent change in consumption models. This change concerns several aspects: awareness of food system issues, employment of socially and environmentally conscious decision-making, food choices based on different characteristics than nutritional ones i.e. origin of food, how it’s produced, and who’s producing it. In this frame the ‘consumption choices’ and the ‘interests of the citizen’ become one part of the others. The figure of the ‘Citizen Consumer’ is born, a responsible and ethically motivated individual to change his lifestyle, achieving the goal of sustainable consumption. Simultaneously the branding, that before was guarantee of the product quality, today is questioned. In order to meet these needs, Agri-Food companies are developing specific product lines that follow two main philosophies: ‘Back to basics’ and ‘Less is more’. However, the issue of ethical behavior does not seem to find an adequate on market offer. Most likely due to a lack of attention on the communication strategy used, very often based on market logic and rarely on ethical one. The label in its classic concept of ‘clean labeling’ can no longer be the only instrument through which to convey product information and its evolution towards a concept of ‘clear label’ is necessary to embrace ethical and transparent concepts in progress the process of democratization of the Food System. The implementation of a voluntary traceability path, relying on the technological models of the Internet of Things or Industry 4.0, would enable the Agri-Food Supply Chain to collect data that, if properly treated, could satisfy the information need of consumers. A change of approach is therefore proposed towards Agri-Food traceability that is no longer intended as a tool to be used to respond to the legislator, but rather as a promotional tool useful to tell the company in a transparent manner and then reach the slice of the market of food citizens. The use of mobile technology can also facilitate this information transfer. However, in order to guarantee maximum efficiency, an appropriate communication model based on the ethical communication principles should be used, which aims to overcome the pipeline communication model, to offer the listener a new way of telling the food product, based on real data collected through processes traceability. The Citizen Consumer is therefore placed at the center of the new model of communication in which he has the opportunity to choose what to know and how. The new label creates a virtual access point capable of telling the product according to different point of views, following the personal interests and offering the possibility to give several content modalities to support different situations and usability.

Keywords: agri food traceability, agri-food transparency, clear label, food system, internet of things

Procedia PDF Downloads 121

Authors: Andrey Cherdantsev, Mikhail Cherdantsev, Sergey Isaenkov, Dmitriy Markovich

Abstract:

In annular gas-liquid flow, liquid flows as a film along pipe walls sheared by high-velocity gas stream. Film surface is covered by large-scale disturbance waves which affect pressure drop and heat transfer in the system and are necessary for entrainment of liquid droplets from film surface into the core of gas stream. Disturbance waves are a highly complex and their properties are affected by numerous parameters. One of such aspects is flow development, i.e., change of flow properties with the distance from the inlet. In the present work, this question is studied using brightness-based laser-induced fluorescence (BBLIF) technique. This method enables one to perform simultaneous measurements of local film thickness in large number of points with high sampling frequency. In the present experiments first 50 cm of upward and downward annular flow in a vertical pipe of 11.7 mm i.d. is studied with temporal resolution of 10 kHz and spatial resolution of 0.5 mm. Thus, spatiotemporal evolution of film surface can be investigated, including scenarios of formation, acceleration and coalescence of disturbance waves. The behaviour of disturbance waves' velocity depending on phases flow rates and downstream distance was investigated. Besides measuring the waves properties, the goal of the work was to investigate the interrelation between disturbance waves properties and integral characteristics of the flow such as interfacial shear stress and flow rate of dispersed phase. In particular, it was shown that the initial acceleration of disturbance waves, defined by the value of shear stress, linearly decays with downstream distance. This lack of acceleration which may even lead to deceleration is related to liquid entrainment. Flow rate of disperse phase linearly grows with downstream distance. During entrainment events, liquid is extracted directly from disturbance waves, reducing their mass, area of interaction to the gas shear and, hence, velocity. Passing frequency of disturbance waves at each downstream position was measured automatically with a new algorithm of identification of characteristic lines of individual disturbance waves. Scenarios of coalescence of individual disturbance waves were identified. Transition from initial high-frequency Kelvin-Helmholtz waves appearing at the inlet to highly nonlinear disturbance waves with lower frequency was studied near the inlet using 3D realisation of BBLIF method in the same cylindrical channel and in a rectangular duct with cross-section of 5 mm by 50 mm. It was shown that the initial waves are generally two-dimensional but are promptly broken into localised three-dimensional wavelets. Coalescence of these wavelets leads to formation of quasi two-dimensional disturbance waves. Using cross-correlation analysis, loss and restoration of two-dimensionality of film surface with downstream distance were studied quantitatively. It was shown that all the processes occur closer to the inlet at higher gas velocities.

Keywords: annular flow, disturbance waves, entrainment, flow development

Procedia PDF Downloads 221

Authors: Marcus Goh Tian Xi, Alicia Chua Si Wen, Eunice Gan Ghee Wu, Helen Bound, Lee Liang Ying, Albert Lee

Abstract:

The Map of Dialogical Inquiry provides a conceptual look at the underlying nature of future-oriented skills. According to the Map, learning is learner-oriented, with conversational time shifted from teachers to learners, who play a strong role in deciding what and how they learn. For example, in courses operating on the principles of Dialogical Inquiry, learners were able to leave the classroom with a deeper understanding of the topic, broader exposure to differing perspectives, and stronger critical thinking capabilities, compared to traditional approaches to teaching. Despite its contributions to learning, the Map is grounded in a qualitative approach both in its development and its application for providing feedback to learners and educators. Studies hinge on openended responses by Map users, which can be time consuming and resource intensive. The present research is motivated by this gap in practicality by aiming to develop and validate a quantitative measure of the Map. In addition, a quantifiable measure may also strengthen applicability by making learning experiences trackable and comparable. The Map outlines eight learning aspects that learners should holistically engage. This research focuses on the Analysing aspect of learning. According to the Map, Analysing has four key components: liking or engaging in logic, using interpretative lenses, seeking patterns, and critiquing and deconstructing. Existing scales of constructs (e.g., critical thinking, rationality) related to these components were identified so that the current scale could adapt items from. Specifically, items were phrased beginning with an “I”, followed by an action phrase, to fulfil the purpose of assessing learners' engagement with Analysing either in general or in classroom contexts. Paralleling standard scale development procedure, the 26-item Analysing scale was administered to 330 participants alongside existing scales with varying levels of association to Analysing, to establish construct validity. Subsequently, the scale was refined and its dimensionality, reliability, and validity were determined. Confirmatory factor analysis (CFA) revealed if scale items loaded onto the four factors corresponding to the components of Analysing. To refine the scale, items were systematically removed via an iterative procedure, according to their factor loadings and results of likelihood ratio tests at each step. Eight items were removed this way. The Analysing scale is better conceptualised as unidimensional, rather than comprising the four components identified by the Map, for three reasons: 1) the covariance matrix of the model specified for the CFA was not positive definite, 2) correlations among the four factors were high, and 3) exploratory factor analyses did not yield an easily interpretable factor structure of Analysing. Regarding validity, since the Analysing scale had higher correlations with conceptually similar scales than conceptually distinct scales, with minor exceptions, construct validity was largely established. Overall, satisfactory reliability and validity of the scale suggest that the current procedure can result in a valid and easy-touse measure for each aspect of the Map.

Keywords: analytical thinking, dialogical inquiry, education, lifelong learning, pedagogy, scale development

Procedia PDF Downloads 62

Authors: Ja-Keoung Koo, Kensuke Nakamura, Hyohun Kim, Dongwha Shin, Yeonseok Kim, Ji-Su Ahn, Byung-Woo Hong

Abstract:

The machine learning techniques based on a convolutional neural network (CNN) have been actively developed and successfully applied to a variety of image analysis tasks including reconstruction, noise reduction, resolution enhancement, segmentation, motion estimation, object recognition. The classical visual information processing that ranges from low level tasks to high level ones has been widely developed in the deep learning framework. It is generally considered as a challenging problem to derive visual interpretation from high dimensional imagery data. A CNN is a class of feed-forward artificial neural network that usually consists of deep layers the connections of which are established by a series of non-linear operations. The CNN architecture is known to be shift invariant due to its shared weights and translation invariance characteristics. However, it is often computationally intractable to optimize the network in particular with a large number of convolution layers due to a large number of unknowns to be optimized with respect to the training set that is generally required to be large enough to effectively generalize the model under consideration. It is also necessary to limit the size of convolution kernels due to the computational expense despite of the recent development of effective parallel processing machinery, which leads to the use of the constantly small size of the convolution kernels throughout the deep CNN architecture. However, it is often desired to consider different scales in the analysis of visual features at different layers in the network. Thus, we propose a CNN model where different sizes of the convolution kernels are applied at each layer based on the random projection. We apply random filters with varying sizes and associate the filter responses with scalar weights that correspond to the standard deviation of the random filters. We are allowed to use large number of random filters with the cost of one scalar unknown for each filter. The computational cost in the back-propagation procedure does not increase with the larger size of the filters even though the additional computational cost is required in the computation of convolution in the feed-forward procedure. The use of random kernels with varying sizes allows to effectively analyze image features at multiple scales leading to a better generalization. The robustness and effectiveness of the proposed CNN based on random kernels are demonstrated by numerical experiments where the quantitative comparison of the well-known CNN architectures and our models that simply replace the convolution kernels with the random filters is performed. The experimental results indicate that our model achieves better performance with less number of unknown weights. The proposed algorithm has a high potential in the application of a variety of visual tasks based on the CNN framework. Acknowledgement—This work was supported by the MISP (Ministry of Science and ICT), Korea, under the National Program for Excellence in SW (20170001000011001) supervised by IITP, and NRF-2014R1A2A1A11051941, NRF2017R1A2B4006023.

Keywords: deep learning, convolutional neural network, random kernel, random projection, dimensionality reduction, object recognition

Procedia PDF Downloads 254

Authors: Hilal T. Sasmazel, Seda Surucu

Abstract:

Skin tissue engineering is a promising field for the treatment of skin defects using scaffolds. This approach involves the use of living cells and biomaterials to restore, maintain, or regenerate tissues and organs in the body by providing; (i) larger surface area for cell attachment, (ii) proper porosity for cell colonization and cell to cell interaction, and (iii) 3-dimensionality at macroscopic scale. Recent studies on this area mainly focus on fabrication of scaffolds that can closely mimic the natural extracellular matrix (ECM) for creation of tissue specific niche-like environment at the subcellular scale. Scaffolds designed as ECM-like architectures incorporating into the host with minimal scarring/pain and facilitate angiogenesis. This study is related to combining of synthetic PCL and natural chitosan polymers to form 3D PCL/Chitosan core-shell structures for skin tissue engineering applications. Amongst the polymers used in tissue engineering, natural polymer chitosan and synthetic polymer poly(ε-caprolactone) (PCL) are widely preferred in the literature. Chitosan has been among researchers for a very long time because of its superior biocompatibility and structural resemblance to the glycosaminoglycan of bone tissue. However, the low mechanical flexibility and limited biodegradability properties reveals the necessity of using this polymer in a composite structure. On the other hand, PCL is a versatile polymer due to its low melting point (60°C), ease of processability, degradability with non-enzymatic processes (hydrolysis) and good mechanical properties. Nevertheless, there are also several disadvantages of PCL such as its hydrophobic structure, limited bio-interaction and susceptibility to bacterial biodegradation. Therefore, it became crucial to use both of these polymers together as a hybrid material in order to overcome the disadvantages of both polymers and combine advantages of those. The scaffolds here were fabricated by using electrospinning technique and the characterizations of the samples were done by contact angle (CA) measurements, scanning electron microscopy (SEM), transmission electron microscopy (TEM) and X-Ray Photoelectron spectroscopy (XPS). Additionally, gas permeability test, mechanical test, thickness measurement and PBS absorption and shrinkage tests were performed for all type of scaffolds (PCL, chitosan and PCL/chitosan core-shell). By using ImageJ launcher software program (USA) from SEM photographs the average inter-fiber diameter values were calculated as 0.717±0.198 µm for PCL, 0.660±0.070 µm for chitosan and 0.412±0.339 µm for PCL/chitosan core-shell structures. Additionally, the average inter-fiber pore size values exhibited decrease of 66.91% and 61.90% for the PCL and chitosan structures respectively, compare to PCL/chitosan core-shell structures. TEM images proved that homogenous and continuous bead free core-shell fibers were obtained. XPS analysis of the PCL/chitosan core-shell structures exhibited the characteristic peaks of PCL and chitosan polymers. Measured average gas permeability value of produced PCL/chitosan core-shell structure was determined 2315±3.4 g.m-2.day-1. In the future, cell-material interactions of those developed PCL/chitosan core-shell structures will be carried out with L929 ATCC CCL-1 mouse fibroblast cell line. Standard MTT assay and microscopic imaging methods will be used for the investigation of the cell attachment, proliferation and growth capacities of the developed materials.

Keywords: chitosan, coaxial electrospinning, core-shell, PCL, tissue scaffold

Procedia PDF Downloads 452

Authors: Ainura Tursunalieva, Irene Hudson

Abstract:

Continuous improvement of both the quality and safety of health care is an important goal in Australia and internationally. The intensive care unit (ICU) receives patients with a wide variety of and severity of illnesses. Accurately identifying patients at risk of developing complications or dying is crucial to increasing healthcare efficiency. Thus, it is essential for clinicians and researchers to have a robust framework capable of evaluating the risk profile of a patient. ICU scoring systems provide such a framework. The Acute Physiology and Chronic Health Evaluation III and the Simplified Acute Physiology Score II are ICU scoring systems frequently used for assessing the severity of acute illness. These scoring systems collect multiple risk factors for each patient including physiological measurements then render the assessment outcomes of individual risk factors into a single numerical value. A higher score is related to a more severe patient condition. Furthermore, the Mortality Probability Model II uses logistic regression based on independent risk factors to predict a patient’s probability of mortality. An important overlooked limitation of SAPS II and MPM II is that they do not, to date, include interaction terms between a patient’s vital signs. This is a prominent oversight as it is likely there is an interplay among vital signs. The co-existence of certain conditions may pose a greater health risk than when these conditions exist independently. One barrier to including such interaction terms in predictive models is the dimensionality issue as it becomes difficult to use variable selection. We propose an innovative scoring system which takes into account a dependence structure among patient’s vital signs, such as systolic and diastolic blood pressures, heart rate, pulse interval, and peripheral oxygen saturation. Copulas will capture the dependence among normally distributed and skewed variables as some of the vital sign distributions are skewed. The estimated dependence parameter will then be incorporated into the traditional scoring systems to adjust the points allocated for the individual vital sign measurements. The same dependence parameter will also be used to create an alternative copula-based model for predicting a patient’s probability of mortality. The new copula-based approach will accommodate not only a patient’s trajectories of vital signs but also the joint dependence probabilities among the vital signs. We hypothesise that this approach will produce more stable assessments and lead to more time efficient and accurate predictions. We will use two data sets: (1) 250 ICU patients admitted once to the Chui Regional Hospital (Kyrgyzstan) and (2) 37 ICU patients’ agitation-sedation profiles collected by the Hunter Medical Research Institute (Australia). Both the traditional scoring approach and our copula-based approach will be evaluated using the Brier score to indicate overall model performance, the concordance (or c) statistic to indicate the discriminative ability (or area under the receiver operating characteristic (ROC) curve), and goodness-of-fit statistics for calibration. We will also report discrimination and calibration values and establish visualization of the copulas and high dimensional regions of risk interrelating two or three vital signs in so-called higher dimensional ROCs.

Keywords: copula, intensive unit scoring system, ROC curves, vital sign dependence

Procedia PDF Downloads 123

Authors: Hyohun Kim, Dongwha Shin, Yeonseok Kim, Ji-Su Ahn, Kensuke Nakamura, Dongeun Choi, Byung-Woo Hong

Abstract:

Artefacts are commonly encountered in the imaging process of clinical computed tomography (CT) where the artefact refers to any systematic discrepancy between the reconstructed observation and the true attenuation coefficient of the object. It is known that CT images are inherently more prone to artefacts due to its image formation process where a large number of independent detectors are involved, and they are assumed to yield consistent measurements. There are a number of different artefact types including noise, beam hardening, scatter, pseudo-enhancement, motion, helical, ring, and metal artefacts, which cause serious difficulties in reading images. Thus, it is desired to remove nuisance factors from the degraded image leaving the fundamental intrinsic information that can provide better interpretation of the anatomical and pathological characteristics. However, it is considered as a difficult task due to the high dimensionality and variability of data to be recovered, which naturally motivates the use of machine learning techniques. We propose an image restoration algorithm based on the deep neural network framework where the denoising auto-encoders are stacked building multiple layers. The denoising auto-encoder is a variant of a classical auto-encoder that takes an input data and maps it to a hidden representation through a deterministic mapping using a non-linear activation function. The latent representation is then mapped back into a reconstruction the size of which is the same as the size of the input data. The reconstruction error can be measured by the traditional squared error assuming the residual follows a normal distribution. In addition to the designed loss function, an effective regularization scheme using residual-driven dropout determined based on the gradient at each layer. The optimal weights are computed by the classical stochastic gradient descent algorithm combined with the back-propagation algorithm. In our algorithm, we initially decompose an input image into its intrinsic representation and the nuisance factors including artefacts based on the classical Total Variation problem that can be efficiently optimized by the convex optimization algorithm such as primal-dual method. The intrinsic forms of the input images are provided to the deep denosing auto-encoders with their original forms in the training phase. In the testing phase, a given image is first decomposed into the intrinsic form and then provided to the trained network to obtain its reconstruction. We apply our algorithm to the restoration of the corrupted CT images by the artefacts. It is shown that our algorithm improves the readability and enhances the anatomical and pathological properties of the object. The quantitative evaluation is performed in terms of the PSNR, and the qualitative evaluation provides significant improvement in reading images despite degrading artefacts. The experimental results indicate the potential of our algorithm as a prior solution to the image interpretation tasks in a variety of medical imaging applications. This work was supported by the MISP(Ministry of Science and ICT), Korea, under the National Program for Excellence in SW (20170001000011001) supervised by the IITP(Institute for Information and Communications Technology Promotion).

Keywords: auto-encoder neural network, CT image artefact, deep learning, intrinsic image representation, noise reduction, total variation

Procedia PDF Downloads 164

Authors: Jeremie Coullon, Robert J. Webber

Abstract:

We introduce a Markov chain Monte Carlo (MCMC) sam-pler for infinite-dimensional inverse problems. Our sam-pler is based on the affine invariant ensemble sampler, which uses interacting walkers to adapt to the covariance structure of the target distribution. We extend this ensem-ble sampler for the first time to infinite-dimensional func-tion spaces, yielding a highly efficient gradient-free MCMC algorithm. Because our ensemble sampler does not require gradients or posterior covariance estimates, it is simple to implement and broadly applicable. In many Bayes-ian inverse problems, Markov chain Monte Carlo (MCMC) meth-ods are needed to approximate distributions on infinite-dimensional function spaces, for example, in groundwater flow, medical imaging, and traffic flow. Yet designing efficient MCMC methods for function spaces has proved challenging. Recent gradi-ent-based MCMC methods preconditioned MCMC methods, and SMC methods have improved the computational efficiency of functional random walk. However, these samplers require gradi-ents or posterior covariance estimates that may be challenging to obtain. Calculating gradients is difficult or impossible in many high-dimensional inverse problems involving a numerical integra-tor with a black-box code base. Additionally, accurately estimating posterior covariances can require a lengthy pilot run or adaptation period. These concerns raise the question: is there a functional sampler that outperforms functional random walk without requir-ing gradients or posterior covariance estimates? To address this question, we consider a gradient-free sampler that avoids explicit covariance estimation yet adapts naturally to the covariance struc-ture of the sampled distribution. This sampler works by consider-ing an ensemble of walkers and interpolating and extrapolating between walkers to make a proposal. This is called the affine in-variant ensemble sampler (AIES), which is easy to tune, easy to parallelize, and efficient at sampling spaces of moderate dimen-sionality (less than 20). The main contribution of this work is to propose a functional ensemble sampler (FES) that combines func-tional random walk and AIES. To apply this sampler, we first cal-culate the Karhunen–Loeve (KL) expansion for the Bayesian prior distribution, assumed to be Gaussian and trace-class. Then, we use AIES to sample the posterior distribution on the low-wavenumber KL components and use the functional random walk to sample the posterior distribution on the high-wavenumber KL components. Alternating between AIES and functional random walk updates, we obtain our functional ensemble sampler that is efficient and easy to use without requiring detailed knowledge of the target dis-tribution. In past work, several authors have proposed splitting the Bayesian posterior into low-wavenumber and high-wavenumber components and then applying enhanced sampling to the low-wavenumber components. Yet compared to these other samplers, FES is unique in its simplicity and broad applicability. FES does not require any derivatives, and the need for derivative-free sam-plers has previously been emphasized. FES also eliminates the requirement for posterior covariance estimates. Lastly, FES is more efficient than other gradient-free samplers in our tests. In two nu-merical examples, we apply FES to challenging inverse problems that involve estimating a functional parameter and one or more scalar parameters. We compare the performance of functional random walk, FES, and an alternative derivative-free sampler that explicitly estimates the posterior covariance matrix. We conclude that FES is the fastest available gradient-free sampler for these challenging and multimodal test problems.

Keywords: Bayesian inverse problems, Markov chain Monte Carlo, infinite-dimensional inverse problems, dimensionality reduction

Procedia PDF Downloads 119