Search results for: optimisation algorithms

Authors: Bastien Batardière, Joon Kwon

Abstract:

For finite-sum optimization, variance-reduced gradient methods (VR) compute at each iteration the gradient of a single function (or of a mini-batch), and yet achieve faster convergence than SGD thanks to a carefully crafted lower-variance stochastic gradient estimator that reuses past gradients. Another important line of research of the past decade in continuous optimization is the adaptive algorithms such as AdaGrad, that dynamically adjust the (possibly coordinate-wise) learning rate to past gradients and thereby adapt to the geometry of the objective function. Variants such as RMSprop and Adam demonstrate outstanding practical performance that have contributed to the success of deep learning. In this work, we present AdaLVR, which combines the AdaGrad algorithm with loopless variance-reduced gradient estimators such as SAGA or L-SVRG that benefits from a straightforward construction and a streamlined analysis. We assess that AdaLVR inherits both good convergence properties from VR methods and the adaptive nature of AdaGrad: in the case of L-smooth convex functions we establish a gradient complexity of O(n + (L + √ nL)/ε) without prior knowledge of L. Numerical experiments demonstrate the superiority of AdaLVR over state-of-the-art methods. Moreover, we empirically show that the RMSprop and Adam algorithm combined with variance-reduced gradients estimators achieve even faster convergence.

Keywords: convex optimization, variance reduction, adaptive algorithms, loopless

Procedia PDF Downloads 71

Authors: Nutcha Taneepanichskul, Helen C. Hailes, Mark Miodownik

Abstract:

Plastic waste has emerged as a critical global environmental challenge, primarily driven by the prevalent use of conventional plastics derived from petrochemical refining and manufacturing processes in modern packaging. While these plastics serve vital functions, their persistence in the environment post-disposal poses significant threats to ecosystems. Addressing this issue necessitates approaches, one of which involves the development of biodegradable plastics designed to degrade under controlled conditions, such as industrial composting facilities. It is imperative to note that compostable plastics are engineered for degradation within specific environments and are not suited for uncontrolled settings, including natural landscapes and aquatic ecosystems. The full benefits of compostable packaging are realized when subjected to industrial composting, preventing environmental contamination and waste stream pollution. Therefore, effective sorting technologies are essential to enhance composting rates for these materials and diminish the risk of contaminating recycling streams. In this study, it leverage hyperspectral imaging technology (HSI) coupled with advanced machine learning algorithms to accurately identify various types of plastics, encompassing conventional variants like Polyethylene terephthalate (PET), Polypropylene (PP), Low density polyethylene (LDPE), High density polyethylene (HDPE) and biodegradable alternatives such as Polybutylene adipate terephthalate (PBAT), Polylactic acid (PLA), and Polyhydroxyalkanoates (PHA). The dataset is partitioned into three subsets: a training dataset comprising uncontaminated conventional and biodegradable plastics, a validation dataset encompassing contaminated plastics of both types, and a testing dataset featuring real-world packaging items in both pristine and contaminated states. Five distinct machine learning algorithms, namely Partial Least Squares Discriminant Analysis (PLS-DA), Support Vector Machine (SVM), Convolutional Neural Network (CNN), Logistic Regression, and Decision Tree Algorithm, were developed and evaluated for their classification performance. Remarkably, the Logistic Regression and CNN model exhibited the most promising outcomes, achieving a perfect accuracy rate of 100% for the training and validation datasets. Notably, the testing dataset yielded an accuracy exceeding 80%. The successful implementation of this sorting technology within recycling and composting facilities holds the potential to significantly elevate recycling and composting rates. As a result, the envisioned circular economy for plastics can be established, thereby offering a viable solution to mitigate plastic pollution.

Keywords: biodegradable plastics, sorting technology, hyperspectral imaging technology, machine learning algorithms

Procedia PDF Downloads 79

Authors: Taysir Soliman

Abstract:

One of the current popular clustering techniques is Spectral Clustering (SC) because of its advantages over conventional approaches such as hierarchical clustering, k-means, etc. and other techniques as well. However, one of the disadvantages of SC is the time consuming process because it requires computing the eigenvectors. In the past to overcome this disadvantage, a number of attempts have been proposed such as the Power Iteration Clustering (PIC) technique, which is one of versions from SC; some of PIC advantages are: 1) its scalability and efficiency, 2) finding one pseudo-eigenvectors instead of computing eigenvectors, and 3) linear combination of the eigenvectors in linear time. However, its worst disadvantage is an inter-class collision problem because it used only one pseudo-eigenvectors which is not enough. Previous researchers developed Deflation-based Power Iteration Clustering (DPIC) to overcome problems of PIC technique on inter-class collision with the same efficiency of PIC. In this paper, we developed Parallel DPIC (PDPIC) to improve the time and memory complexity which is run on apache spark framework using sparse matrix. To test the performance of PDPIC, we compared it to SC, ESCG, ESCALG algorithms on four small graph benchmark datasets and nine large graph benchmark datasets, where PDPIC proved higher accuracy and better time consuming than other compared algorithms.

Keywords: spectral clustering, power iteration clustering, deflation-based power iteration clustering, Apache spark, large graph

Procedia PDF Downloads 189

Authors: Alireza Gholami, Amir H. D. Markazi

Abstract:

In this paper, an observer-based direct adaptive fuzzy sliding mode (OAFSM) algorithm is proposed. In the proposed algorithm, the zero-input dynamics of the plant could be unknown. The input connection matrix is used to combine the sliding surfaces of individual subsystems, and an adaptive fuzzy algorithm is used to estimate an equivalent sliding mode control input directly. The fuzzy membership functions, which were determined by time consuming try and error processes in previous works, are adjusted by adaptive algorithms. The other advantage of the proposed controller is that the input gain matrix is not limited to be diagonal, i.e. the plant could be over/under actuated provided that controllability and observability are preserved. An observer is constructed to directly estimate the state tracking error, and the nonlinear part of the observer is constructed by an adaptive fuzzy algorithm. The main advantage of the proposed observer is that, the measured outputs is not limited to the first entry of a canonical-form state vector. The closed-loop stability of the proposed method is proved using a Lyapunov-based approach. The proposed method is applied numerically on a multi-link robot manipulator, which verifies the performance of the closed-loop control. Moreover, the performance of the proposed algorithm is compared with some conventional control algorithms.

Keywords: adaptive algorithm, fuzzy systems, membership functions, observer

Procedia PDF Downloads 206

Authors: J. Niemisto, A. Nissinen, S. Soimakallio

Abstract:

Aviation causes carbon dioxide (CO2) emissions and other climate forcers which increase the contribution of aviation on climate change. Aviation industry and number of air travellers are constantly increasing. Aviation industry has an ambitious goal to strongly cut net CO2 emissions. Modern fleet, alternative jet fuels technologies and route optimisation are important technological tools in the emission reduction. Faster approaches are needed as well. Emission trade systems, voluntary carbon offset compensation schemes and taxation are already in operation. Global scenarios of aviation industry and its greenhouse gas emissions and other climate forcers are discussed in this review study based on literature and other published data. The focus is on the aviation in Nordic countries, but also European and global situation are considered. Different emission reduction technologies and compensation modes are examined. In addition, the role of aviation in a single passenger’s (a Finnish consumer) annual carbon footprint is analysed and a comparison of available emission calculators and carbon offset systems is performed. Long-haul fights have a significant role in a single consumer´s and company´s carbon footprint, but remarkable change in global emission level would need a huge change in attitudes towards flying.

Keywords: aviation, climate change, emissions, environment

Procedia PDF Downloads 211

Authors: I. Stathopoulos, V. Syrgiamiotis, E. Karavasilis, A. Ploussi, I. Nikas, C. Hatzigiorgi, K. Platoni, E. P. Efstathopoulos

Abstract:

Introduction: Brain and central nervous system (CNS) tumors form the second most common group of cancer in children, accounting for 30% of all childhood cancers. MRI is the key imaging technique used for the visualization and management of pediatric brain tumors. Initial characterization of tumors from MRI scans is usually performed via a radiologist’s visual assessment. However, different brain tumor types do not always demonstrate clear differences in visual appearance. Using only conventional MRI to provide a definite diagnosis could potentially lead to inaccurate results, and so histopathological examination of biopsy samples is currently considered to be the gold standard for obtaining definite diagnoses. Machine learning is defined as the study of computational algorithms that can use, complex or not, mathematical relationships and patterns from empirical and scientific data to make reliable decisions. Concerning the above, machine learning techniques could provide effective and accurate ways to automate and speed up the analysis and diagnosis for medical images. Machine learning applications in radiology are or could potentially be useful in practice for medical image segmentation and registration, computer-aided detection and diagnosis systems for CT, MR or radiography images and functional MR (fMRI) images for brain activity analysis and neurological disease diagnosis. Purpose: The objective of this study is to provide an automated tool, which may assist in the imaging evaluation and classification of brain neoplasms in pediatric patients by determining the glioma type, grade and differentiating between different brain tissue types. Moreover, a future purpose is to present an alternative way of quick and accurate diagnosis in order to save time and resources in the daily medical workflow. Materials and Methods: A cohort, of 80 pediatric patients with a diagnosis of posterior fossa tumor, was used: 20 ependymomas, 20 astrocytomas, 20 medulloblastomas and 20 healthy children. The MR sequences used, for every single patient, were the following: axial T1-weighted (T1), axial T2-weighted (T2), FluidAttenuated Inversion Recovery (FLAIR), axial diffusion weighted images (DWI), axial contrast-enhanced T1-weighted (T1ce). From every sequence only a principal slice was used that manually traced by two expert radiologists. Image acquisition was carried out on a GE HDxt 1.5-T scanner. The images were preprocessed following a number of steps including noise reduction, bias-field correction, thresholding, coregistration of all sequences (T1, T2, T1ce, FLAIR, DWI), skull stripping, and histogram matching. A large number of features for investigation were chosen, which included age, tumor shape characteristics, image intensity characteristics and texture features. After selecting the features for achieving the highest accuracy using the least number of variables, four machine learning classification algorithms were used: k-Nearest Neighbour, Support-Vector Machines, C4.5 Decision Tree and Convolutional Neural Network. The machine learning schemes and the image analysis are implemented in the WEKA platform and MatLab platform respectively. Results-Conclusions: The results and the accuracy of images classification for each type of glioma by the four different algorithms are still on process.

Keywords: image classification, machine learning algorithms, pediatric MRI, pediatric oncology

Procedia PDF Downloads 149

Authors: Mohamed Al-Badrashiny, Abdelghani Bellaachia

Abstract:

This paper presents a classifier ensemble approach for predicting the survivability of the breast cancer patients using the latest database version of the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute. The system consists of two main components; features selection and classifier ensemble components. The features selection component divides the features in SEER database into four groups. After that it tries to find the most important features among the four groups that maximizes the weighted average F-score of a certain classification algorithm. The ensemble component uses three different classifiers, each of which models different set of features from SEER through the features selection module. On top of them, another classifier is used to give the final decision based on the output decisions and confidence scores from each of the underlying classifiers. Different classification algorithms have been examined; the best setup found is by using the decision tree, Bayesian network, and Na¨ıve Bayes algorithms for the underlying classifiers and Na¨ıve Bayes for the classifier ensemble step. The system outperforms all published systems to date when evaluated against the exact same data of SEER (period of 1973-2002). It gives 87.39% weighted average F-score compared to 85.82% and 81.34% of the other published systems. By increasing the data size to cover the whole database (period of 1973-2014), the overall weighted average F-score jumps to 92.4% on the held out unseen test set.

Keywords: classifier ensemble, breast cancer survivability, data mining, SEER

Procedia PDF Downloads 328

Authors: Yadgar Tayfur, Antony Darby, Tim Ibell, Mark Evernden, John Orr

Abstract:

Fabric form-work is a technique to cast concrete structures with a great advantage of saving concrete material of up to 40%. This technique is particularly associated with the optimized concrete structures that usually have smaller cross-section dimensions than equivalent prismatic members. However, this can make the structural system produced from these members prone to smaller serviceability safety margins. Therefore, it is very important to understand the serviceability issue of non-prismatic concrete structures. In this paper, an analytical computer-based model to optimize concrete beams and to predict load-deflection behaviour of both prismatic and non-prismatic concrete beams is presented. The model was developed based on the method of sectional analysis and integration of curvatures. Results from the analytical model were compared to load-deflection behaviour of a number of beams with different geometric and material properties from other researchers. The results of the comparison show that the analytical program can accurately predict the load-deflection response of concrete beams with medium reinforcement ratios. However, it over-estimates deflection values for lightly reinforced specimens. Finally, the analytical program acceptably predicted load-deflection behaviour of on-prismatic concrete beams.

Keywords: fabric-formed concrete, continuous beams, optimisation, serviceability

Procedia PDF Downloads 372

Authors: Samuel Knox, Verdon Crann, Peyman Amiri, William Crowther

Abstract:

There exists a shortage of open-source firmware for allowing researchers to focus on implementing high-level planning and control strategies for multi aerial robotic systems in simulation and experiment. Within this body of work, practical firmware is presented, which performs all supplementary tasks, including communications, pre and post-experiment procedures, and emergency safety measures. This allows researchers to implement high-level planning and control algorithms for path planning, traffic management, flight formation and swarming of aerial robots. The framework is built in Python using the MAVSDK library, which is compatible with flight controllers running PX4 firmware and onboard computers based on Linux. Communication is performed using Wi-Fi and the MQTT protocol, currently implemented using a centralized broker. Finally, a graphical user interface (GUI) has been developed to send general commands and monitor the agents. This framework enables researchers to prepare customized planning and control algorithms in a modular manner. Studies can be performed experimentally and in simulation using PX4 software in the loop (SITL) and the Gazebo simulator. An example experimental use case of the framework is presented using novel distributed planning and control strategies. The demonstration is performed using off-the-shelf components and minimal setup.

Keywords: aerial robotics, distributed framework, experimental, planning and control

Procedia PDF Downloads 113

Authors: Feng Tao, Han Ye, Shaoyi Liao

Abstract:

City constructions collapse after earthquake and people will be buried under ruins. Search and rescue should be conducted as soon as possible to save them. Therefore, according to the complicated environment, irregular aftershocks and rescue allow of no delay, a kind of target localization method based on RSSI (Received Signal Strength Indication) is proposed in this article. The target localization technology based on RSSI with the features of low cost and low complexity has been widely applied to nodes localization in WSN (Wireless Sensor Networks). Based on the theory of RSSI transmission and the environment impact to RSSI, this article conducts the experiments in five scenes, and multiple filtering algorithms are applied to original RSSI value in order to establish the signal propagation model with minimum test error respectively. Target location can be calculated from the distance, which can be estimated from signal propagation model, through improved centroid algorithm. Result shows that the localization technology based on RSSI is suitable for large-scale nodes localization. Among filtering algorithms, mixed filtering algorithm (average of average, median and Gaussian filtering) performs better than any other single filtering algorithm, and by using the signal propagation model, the minimum error of distance between known nodes and target node in the five scene is about 3.06m.

Keywords: signal propagation model, centroid algorithm, localization, mixed filtering, RSSI

Procedia PDF Downloads 300

Authors: Roya Ahmadi Ahangar, Hamid Madadyari

Abstract:

The seeker optimization algorithm (SOA) is increasingly gaining popularity among the researchers society due to its effectiveness in solving some real-world optimization problems. This paper provides the load-frequency control method based on the SOA for removing oscillations in the power system. A three-zone power system includes a thermal zone, a hydraulic zone and a wind zone equipped with robust proportional-integral-differential (PID) controllers. The result of simulation indicates that load-frequency changes in the wind zone for the multi-zone system are damped in a short period of time. Meanwhile, in the oscillation period, the oscillations amplitude is not significant. The result of simulation emphasizes that the PID controller designed using the seeker optimization algorithm has a robust function and a better performance for oscillations damping compared to the traditional PID controller. The proposed controller’s performance has been compared to the performance of PID controller regulated with Particle Swarm Optimization (PSO) and. Genetic Algorithm (GA) and Artificial Bee Colony (ABC) algorithms in order to show the superior capability of the proposed SOA in regulating the PID controller. The simulation results emphasize the better performance of the optimized PID controller based on SOA compared to the PID controller optimized with PSO, GA and ABC algorithms.

Keywords: load-frequency control, multi zone, robust PID controller, wind generation

Procedia PDF Downloads 304

Authors: A. Degale Desta, Cheng Jian

Abstract:

Deep learning application in computer vision is rapidly advancing, giving it the ability to monitor the public and quickly identify potentially anomalous behaviour from crowd scenes. Therefore, the purpose of the current work is to improve the performance of safety of people in crowd events from panic behaviour through introducing the innovative idea of Aggregation of Ensembles (AOE), which makes use of the pre-trained ConvNets and a pool of classifiers to find anomalies in video data with packed scenes. According to the theory of algorithms that applied K-means, KNN, CNN, SVD, and Faster-CNN, YOLOv5 architectures learn different levels of semantic representation from crowd videos; the proposed approach leverages an ensemble of various fine-tuned convolutional neural networks (CNN), allowing for the extraction of enriched feature sets. In addition to the above algorithms, a long short-term memory neural network to forecast future feature values and a handmade feature that takes into consideration the peculiarities of the crowd to understand human behavior. On well-known datasets of panic situations, experiments are run to assess the effectiveness and precision of the suggested method. Results reveal that, compared to state-of-the-art methodologies, the system produces better and more promising results in terms of accuracy and processing speed.

Keywords: action recognition, computer vision, crowd detecting and tracking, deep learning

Procedia PDF Downloads 161

Authors: Shryas Bhurat, Aryan Vashistha, Sampreet Dinakar Nayak, Ayush Gupta

Abstract:

As the technology is advancing more towards high computational resources, there is a paradigm shift in the usage of these resources to optimize the design process. This paper discusses the usage of ‘Generative Design using Artificial Intelligence’ to build better models that adapt the operations like selection, mutation, and crossover to generate results. The human mind thinks of the simplest approach while designing an object, but the intelligence learns from the past & designs the complex optimized CAD Models. Generative Design takes the boundary conditions and comes up with multiple solutions with iterations to come up with a sturdy design with the most optimal parameter that is given, saving huge amounts of time & resources. The new production techniques that are at our disposal allow us to use additive manufacturing, 3D printing, and other innovative manufacturing techniques to save resources and design artistically engineered CAD Models. Also, this paper discusses the Genetic Algorithm, the Non-Domination technique to choose the right results using biomimicry that has evolved for current habitation for millions of years. The computer uses parametric models to generate newer models using an iterative approach & uses cloud computing to store these iterative designs. The later part of the paper compares the topology optimization technology with Generative Design that is previously being used to generate CAD Models. Finally, this paper shows the performance of algorithms and how these algorithms help in designing resource-efficient models.

Keywords: genetic algorithm, bio mimicry, generative modeling, non-dominant techniques

Procedia PDF Downloads 149

Authors: Shaira L. Kee, Michael Aaron G. Sy, Myles Joshua T. Tan, Hezerul Abdul Karim, Nouar AlDahoul

Abstract:

Skin diseases are considered the fourth most common disease, with melanoma and non-melanoma skin cancer as the most common type of cancer in Caucasians. The alarming increase in Skin Cancer cases shows an urgent need for further research to improve diagnostic methods, as early diagnosis can significantly improve the 5-year survival rate. Machine Learning algorithms for image pattern analysis in diagnosing skin lesions can dramatically increase the accuracy rate of detection and decrease possible human errors. Several studies have shown the diagnostic performance of computer algorithms outperformed dermatologists. However, existing methods still need improvements to reduce diagnostic errors and generate efficient and accurate results. Our paper proposes an ensemble method to classify dermoscopic images into benign and malignant skin lesions. The experiments were conducted using the International Skin Imaging Collaboration (ISIC) image samples. The dataset contains 3,297 dermoscopic images with benign and malignant categories. The results show improvement in performance with an accuracy of 88% and an F1 score of 87%, outperforming other existing models such as support vector machine (SVM), Residual network (ResNet50), EfficientNetB0, EfficientNetB4, and VGG16.

Keywords: deep learning - VGG16 - efficientNet - CNN – ensemble – dermoscopic images - melanoma

Procedia PDF Downloads 81

Authors: Hankyu Lee, Seoung Bum Kim

Abstract:

The issue with imbalance and overlapping in the class distribution becomes important in various applications of data mining. The imbalanced dataset is a special case in classification problems in which the number of observations of one class (i.e., major class) heavily exceeds the number of observations of the other class (i.e., minor class). Overlapped dataset is the case where many observations are shared together between the two classes. Imbalanced and overlapped data can be frequently found in many real examples including fraud and abuse patients in healthcare, quality prediction in manufacturing, text classification, oil spill detection, remote sensing, and so on. The class imbalance and overlap problem is the challenging issue because this situation degrades the performance of most of the standard classification algorithms. In this study, we propose a classification procedure that can effectively handle imbalanced and overlapped datasets by splitting data space into three parts: nonoverlapping, light overlapping, and severe overlapping and applying the classification algorithm in each part. These three parts were determined based on the Hausdorff distance and the margin of the modified support vector machine. An experiments study was conducted to examine the properties of the proposed method and compared it with other classification algorithms. The results showed that the proposed method outperformed the competitors under various imbalanced and overlapped situations. Moreover, the applicability of the proposed method was demonstrated through the experiment with real data.

Keywords: classification, imbalanced data with class overlap, split data space, support vector machine

Procedia PDF Downloads 308

Authors: Chia-Yen Lee, Hao-Jen Wang, Jhih-Hao Lai

Abstract:

In recent years, infrared (IR) imaging has been considered as a potential tool to assess the efficacy of chemotherapy and early detection of breast cancer. Regions of tumor growth with high metabolic rate and angiogenesis phenomenon lead to the high temperatures. Observation of differences between the heat maps in long term is useful to help assess the growth of breast cancer cells and detect breast cancer earlier, wherein the multi-time infrared image alignment technology is a necessary step. Representative feature points detection and matching are essential steps toward the good performance of image registration and quantitative analysis. However, there is no clear boundary on the infrared images and the subject's posture are different for each shot. It cannot adhesive markers on a body surface for a very long period, and it is hard to find anatomic fiducial markers on a body surface. In other words, it’s difficult to detect and match features in an IR sequence images. In this study, automated feature detection and matching algorithms with two type of automatic feature points (i.e., vascular branch points and modified Harris corner) are developed respectively. The preliminary results show that the proposed method could identify the representative feature points on the IR breast images successfully of 98% accuracy and the matching results of 93% accuracy.

Keywords: Harris corner, infrared image, feature detection, registration, matching

Procedia PDF Downloads 304

Authors: M. P. E. E Silva, A. M. Galloway, A. I. Toumpis

Abstract:

An increasing concern exists in the welding industry in terms of faster joining processes. Methods such as the alternation between shielding gases such Ar, CO₂ and He have been able to provide improved penetration of the joint, reduced heat transfer to the workpiece, and increased travel speeds of the welding torch. Nitrogen as a shielding gas is not desirable due to its reactive behavior within the arc plasma, being absorbed by the molten pool during the welding process. Below certain amounts, nitrogen is not harmful. However, the nitrogen threshold is reduced during the solidification of the joint, and if its subsequent desorption is not completed on time, gas entrapment and blowhole formation may occur. The present study expanded the use of the alternating shielding gas method in the gas metal arc welding (GMAW) process by alternately supplying Ar/5%N₂ and He. Improvements were introduced in terms of joint strength and grain refinement. Microstructural characterization findings showed porosity-free welds with reduced inclusion formation while mechanical tests such as tensile and bend tests confirmed the reinforcement of the joint by the addition of nitrogen. Additionally, significant reductions of the final distortion of the workpiece were found after the welding procedure as well as decreased heat affected zones and temperatures of the weld.

Keywords: alternating shielding gas method, GMAW, grain refinement, nitrogen, porosity, mechanical testing

Procedia PDF Downloads 110

Authors: Atefeh Heydari, Mohammadali Tavakoli, Naomie Salim

Abstract:

In recent years, we have witnessed that online reviews are the most important source of customers’ opinion. They are progressively more used by individuals and organisations to make purchase and business decisions. Unfortunately, for the reason of profit or fame, frauds produce deceptive reviews to hoodwink potential customers. Their activities mislead not only potential customers to make appropriate purchasing decisions and organisations to reshape their business, but also opinion mining techniques by preventing them from reaching accurate results. Spam reviews could be divided into two main groups, i.e. multiple and singleton spam reviews. Detecting a singleton spam review that is the only review written by a user ID is extremely challenging due to lack of clue for detection purposes. Singleton spam reviews are very harmful and various features and proofs used in multiple spam reviews detection are not applicable in this case. Current research aims to propose a novel supervised technique to detect singleton spam reviews. To achieve this, various features are proposed in this study and are to be combined with the most appropriate features extracted from literature and employed in a classifier. In order to compare the performance of different classifiers, SVM and naive Bayes classification algorithms were used for model building. The results revealed that SVM was more accurate than naive Bayes and our proposed technique is capable to detect singleton spam reviews effectively.

Keywords: classification algorithms, Naïve Bayes, opinion review spam detection, singleton review spam detection, support vector machine

Procedia PDF Downloads 309

Authors: Emmanuel Okechukwu Nwafor, Folake Olubunmi Akintayo, Denis Alcides Rezende

Abstract:

Decision tree models are versatile and interpretable machine learning algorithms widely used for both classification and regression tasks, which can be related to cities, whether physical or digital. The aim of this research is to assess how well decision tree algorithms can predict trip purposes in Makurdi, Nigeria, while also exploring their connection to the strategic digital city initiative. The research methodology involves formalizing household demographic and trips information datasets obtained from extensive survey process. Modelling and Prediction were achieved using Python Programming Language and the evaluation metrics like R-squared and mean absolute error were used to assess the decision tree algorithm's performance. The results indicate that the model performed well, with accuracies of 84% and 68%, and low MAE values of 0.188 and 0.314, on training and validation data, respectively. This suggests the model can be relied upon for future prediction. The conclusion reiterates that This model will assist decision-makers, including urban planners, transportation engineers, government officials, and commuters, in making informed decisions on transportation planning and management within the framework of a strategic digital city. Its application will enhance the efficiency, sustainability, and overall quality of transportation services in Makurdi, Nigeria.

Keywords: decision tree algorithm, trip purpose, intelligent transport, strategic digital city, travel pattern, sustainable transport

Procedia PDF Downloads 21

Authors: Vanessa Kummer, Maik Meusel

Abstract:

Real estate demand estimates are typically derived from transaction data. However, in regions with excess demand, transactions are driven by supply and therefore do not indicate what people are actually looking for. To estimate the demand for housing in Switzerland, search subscriptions from all important Swiss real estate platforms are used. These data do, however, suffer from missing information—for example, many users do not specify how many rooms they would like or what price they would be willing to pay. In economic analyses, it is often the case that only complete data is used. Usually, however, the proportion of complete data is rather small which leads to most information being neglected. Also, the data might have a strong distortion if it is complete. In addition, the reason that data is missing might itself also contain information, which is however ignored with that approach. An interesting issue is, therefore, if for economic analyses such as the one at hand, there is an added value by using the whole data set with the imputed missing values compared to using the usually small percentage of complete data (baseline). Also, it is interesting to see how different algorithms affect that result. The imputation of the missing data is done using unsupervised learning. Out of the numerous unsupervised learning approaches, the most common ones, such as clustering, principal component analysis, or neural networks techniques are applied. By training the model iteratively on the imputed data and, thereby, including the information of all data into the model, the distortion of the first training set—the complete data—vanishes. In a next step, the performances of the algorithms are measured. This is done by randomly creating missing values in subsets of the data, estimating those values with the relevant algorithms and several parameter combinations, and comparing the estimates to the actual data. After having found the optimal parameter set for each algorithm, the missing values are being imputed. Using the resulting data sets, the next step is to estimate the willingness to pay for real estate. This is done by fitting price distributions for real estate properties with certain characteristics, such as the region or the number of rooms. Based on these distributions, survival functions are computed to obtain the functional relationship between characteristics and selling probabilities. Comparing the survival functions shows that estimates which are based on imputed data sets do not differ significantly from each other; however, the demand estimate that is derived from the baseline data does. This indicates that the baseline data set does not include all available information and is therefore not representative for the entire sample. Also, demand estimates derived from the whole data set are much more accurate than the baseline estimation. Thus, in order to obtain optimal results, it is important to make use of all available data, even though it involves additional procedures such as data imputation.

Keywords: demand estimate, missing-data imputation, real estate, unsupervised learning

Procedia PDF Downloads 285

Authors: Simon Ingelsten, Anton Lundberg, Vijay Shankar, Lars-Olof Landström, Örjan Johansson

Abstract:

The flow in a low consistency disc refiner is simulated with the aim of identifying flow structures possibly being of importance for a future study to optimise the energy efficiency in refining processes. A simplified flow geometry is used, where a single groove of a refiner disc is modelled. Two different fibre models are used to simulate turbulent fibre suspension flow in the groove. The first model is a Bingham viscoplastic fluid model where the fibre suspension is treated as a non-Newtonian fluid with a yield stress. The second model is a new model proposed in a recent study where the suspended fibres effect on flow is accounted for through a modelled orientation distribution function (ODF). Both models yielded similar results with small differences. Certain flow characteristics that were expected and that was found in the literature were identified. Some of these flow characteristics may be of importance in a future process to optimise the refiner geometry to increase the energy efficiency. Further study and a more detailed flow model is; however, needed in order for the simulations to yield results valid for quantitative use in such an optimisation study. An outline of the next steps in such a study is proposed.

Keywords: disc refiner, fibre flow, sustainability, turbulence modelling

Procedia PDF Downloads 407

Authors: Yeliz Karaca, Carlo Cattani

Abstract:

This study focuses on the local Holder exponent as a measure of the function regularity for time series related to finance data. In this study, the attributes of the finance dataset belonging to 13 countries (India, China, Japan, Sweden, France, Germany, Italy, Australia, Mexico, United Kingdom, Argentina, Brazil, USA) located in 5 different continents (Asia, Europe, Australia, North America and South America) have been examined.These countries are the ones mostly affected by the attributes with regard to financial development, covering a period from 2012 to 2017. Our study is concerned with the most important attributes that have impact on the development of finance for the countries identified. Our method is comprised of the following stages: (a) among the multi fractal methods and Brownian motion Holder regularity functions (polynomial, exponential), significant and self-similar attributes have been identified (b) The significant and self-similar attributes have been applied to the Artificial Neuronal Network (ANN) algorithms (Feed Forward Back Propagation (FFBP) and Cascade Forward Back Propagation (CFBP)) (c) the outcomes of classification accuracy have been compared concerning the attributes that have impact on the attributes which affect the countries’ financial development. This study has enabled to reveal, through the application of ANN algorithms, how the most significant attributes are identified within the relevant dataset via the Holder functions (polynomial and exponential function).

Keywords: artificial neural networks, finance data, Holder regularity, multifractals

Procedia PDF Downloads 246

Authors: John-Nelson Ekumah, Selorm Yao-Say Solomon Adade, Mingming Zhong, Yufan Sun, Qiufang Liang, Muhammad Safiullah Virk, Xorlali Nunekpeku, Nana Adwoa Nkuma Johnson, Bridget Ama Kwadzokpui, Xiaofeng Ren

Abstract:

Enhancing starch gel strength and stability is crucial. However, traditional gel property assessment methods are destructive, time-consuming, and resource-intensive. Thus, understanding ethanol treatment effects on kudzu starch gel strength and developing a rapid, nondestructive gel strength assessment method is essential for optimizing the treatment process and ensuring product quality consistency. This study investigated the effects of different ethanol concentrations on the microstructure of kudzu starch gels using a comprehensive microstructural analysis. We also developed a nondestructive method for predicting gel strength and classifying treatment levels using near-infrared (NIR) spectroscopy, and advanced data analytics. Scanning electron microscopy revealed progressive network densification and pore collapse with increasing ethanol concentration, correlating with enhanced mechanical properties. NIR spectroscopy, combined with various variable selection methods (CARS, GA, and UVE) and modeling algorithms (PLS, SVM, and ELM), was employed to develop predictive models for gel strength. The UVE-SVM model demonstrated exceptional performance, with the highest R² values (Rc = 0.9786, Rp = 0.9688) and lowest error rates (RMSEC = 6.1340, RMSEP = 6.0283). Pattern recognition algorithms (PCA, LDA, and KNN) successfully classified gels based on ethanol treatment levels, achieving near-perfect accuracy. This integrated approach provided a multiscale perspective on ethanol-induced starch gel modification, from molecular interactions to macroscopic properties. Our findings demonstrate the potential of NIR spectroscopy, coupled with advanced data analysis, as a powerful tool for rapid, nondestructive quality assessment in starch gel production. This study contributes significantly to the understanding of starch modification processes and opens new avenues for research and industrial applications in food science, pharmaceuticals, and biomaterials.

Keywords: kudzu starch gel, near-infrared spectroscopy, gel strength prediction, support vector machine, pattern recognition algorithms, ethanol treatment

Procedia PDF Downloads 37

Authors: G. Kaur, A. P. Kulkarni, S. Giddey

Abstract:

Production of fuels and chemicals using renewable energy is a promising way for large-scale energy storage and export. Solid oxide electrolysers (SOEs) integrated with renewable source of energy can produce 'Syngas' H₂/CO from H₂O/CO₂ in the desired ratio for further conversion to liquid fuels. As only a waste CO₂ from industrial and power generation processes is utilized in these processes, this approach is CO₂ neutral compared to using fossil fuel feedstock. In addition, the waste heat from industrial processes or heat from solar thermal concentrators can be effectively utilised in SOEs to further reduce the electrical requirements by up to 30% which boosts overall energy efficiency of the process. In this paper, the electrochemical performance of various novel steam/CO₂ reduction electrodes (cathode) would be presented. The efficiency and lifetime degradation data for single cells and a stack would be presented along with the response of cells to variable electrical load input mimicking the intermittent nature of the renewable energy sources. With such optimisation, newly developed electrodes have been tested for 500+ hrs with Faraday efficiency (electricity to fuel conversion efficiency) up to 95%, and thermal efficiency in excess of 70% based upon energy content of the syngas produced.

Keywords: carbon dioxide, steam conversion, electrochemical system, energy storage, fuel production, renewable energy

Procedia PDF Downloads 237

Authors: U. Borasi, Amit Kr. Jain, Rakesh, Piyush Jain

Abstract:

Data collection, data preparation, model training, model evaluation, and deployment are all processes in a typical machine learning workflow. Training data needs to be gathered and organised. This often entails collecting a sizable dataset and cleaning it to remove or correct any inaccurate or missing information. Preparing the data for use in the machine learning model requires pre-processing it after it has been acquired. This often entails actions like scaling or normalising the data, handling outliers, selecting appropriate features, reducing dimensionality, etc. This pre-processed data is then used to train a model on some machine learning algorithm. After the model has been trained, it needs to be assessed by determining metrics like accuracy, precision, and recall, utilising a test dataset. Every time a new model is built, both data pre-processing and model training—two crucial processes in the Machine learning (ML) workflow—must be carried out. Thus, there are various Machine Learning algorithms that can be employed for every single approach to data pre-processing, generating a large set of combinations to choose from. Example: for every method to handle missing values (dropping records, replacing with mean, etc.), for every scaling technique, and for every combination of features selected, a different algorithm can be used. As a result, in order to get the optimum outcomes, these tasks are frequently repeated in different combinations. This paper suggests a simple architecture for organizing this largely produced “combination set of pre-processing steps and algorithms” into an automated workflow which simplifies the task of carrying out all possibilities.

Keywords: machine learning, automation, AUTOML, architecture, operator pool, configuration, scheduler

Procedia PDF Downloads 58

Authors: Chongya Song, Kang Yen, Alexander Pons, Jin Liu

Abstract:

In the field of machine learning, the ensemble has been employed as a common methodology to improve the performance upon multiple base classifiers. However, the true predictions are often canceled out by the false ones during consensus due to a phenomenon called “curse of correlation” which is represented as the strong interferences among the predictions produced by the base classifiers. In addition, the existing practices are still not able to effectively mitigate the problem of imbalanced classification. Based on the analysis on our experiment results, we conclude that the two problems are caused by some inherent deficiencies in the approach of consensus. Therefore, we create an enhanced ensemble algorithm which adopts a designed rank-based chain-mode consensus to overcome the two problems. In order to evaluate the proposed ensemble algorithm, we employ a well-known benchmark data set NSL-KDD (the improved version of dataset KDDCup99 produced by University of New Brunswick) to make comparisons between the proposed and 8 common ensemble algorithms. Particularly, each compared ensemble classifier uses the same 22 base classifiers, so that the differences in terms of the improvements toward the accuracy and reliability upon the base classifiers can be truly revealed. As a result, the proposed rank-based chain-mode consensus is proved to be a more effective ensemble solution than the traditional consensus approach, which outperforms the 8 ensemble algorithms by 20% on almost all compared metrices which include accuracy, precision, recall, F1-score and area under receiver operating characteristic curve.

Keywords: consensus, curse of correlation, imbalance classification, rank-based chain-mode ensemble

Procedia PDF Downloads 138

Authors: Kemka Ihemelandu, Chukwuemeka Ihemelandu

Abstract:

Melanoma remains a public health crisis, with incidence rates increasing rapidly in the past decades. Improving diagnostic accuracy to decrease misdiagnosis using Artificial intelligence (AI) continues to be documented. Unfortunately, unintended racially biased outcomes, a product of lack of diversity in the dataset used, with a noted class imbalance favoring lighter vs. darker skin tone, have increasingly been recognized as a problem.Resulting in noted limitations of the accuracy of the Convolutional neural network (CNN)models. CNN models are prone to biased output due to biases in the dataset used to train them. Our aim in this study was the optimization of convolutional neural network algorithms to mitigate bias in the automated diagnosis of melanoma. We hypothesized that our proposed training algorithms based on a data augmentation method to optimize the diagnostic accuracy of a CNN classifier by generating new training samples from the original ones will reduce bias in the automated diagnosis of melanoma. We applied geometric transformation, including; rotations, translations, scale change, flipping, and shearing. Resulting in a CNN model that provided a modifiedinput data making for a model that could learn subtle racial features. Optimal selection of the momentum and batch hyperparameter increased our model accuracy. We show that our augmented model reduces bias while maintaining accuracy in the automated diagnosis of melanoma.

Keywords: bias, augmentation, melanoma, convolutional neural network

Procedia PDF Downloads 211

Authors: Maxwell Henderson, Tristan Cook, Justin Chan Jin Le, Mark Hodson, YoungJung Chang, John Novak, Daniel Padilha, Nishan Kulatilaka, Ansu Bagchi, Sanjoy Ray, John Kelly

Abstract:

We present a method of using adiabatic quantum optimization (AQO) to solve a mixed integer nonlinear programming (MINLP) problem instance. The MINLP problem is a general form of a set of NP-hard optimization problems that are critical to many business applications. It requires optimizing a set of discrete and continuous variables with nonlinear and potentially nonconvex constraints. Obtaining an exact, optimal solution for MINLP problem instances of non-trivial size using classical computation methods is currently intractable. Current leading algorithms leverage heuristic and divide-and-conquer methods to determine approximate solutions. Creating more accurate and efficient algorithms is an active area of research. Quantum computing (QC) has several theoretical benefits compared to classical computing, through which QC algorithms could obtain MINLP solutions that are superior to current algorithms. AQO is a particular form of QC that could offer more near-term benefits compared to other forms of QC, as hardware development is in a more mature state and devices are currently commercially available from D-Wave Systems Inc. It is also designed for optimization problems: it uses an effect called quantum tunneling to explore all lowest points of an energy landscape where classical approaches could become stuck in local minima. Our work used a novel algorithm formulated for AQO to solve a special type of MINLP problem. The research focused on determining: 1) if the problem is possible to solve using AQO, 2) if it can be solved by current hardware, 3) what the currently achievable performance is, 4) what the performance will be on projected future hardware, and 5) when AQO is likely to provide a benefit over classical computing methods. Two different methods, integer range and 1-hot encoding, were investigated for transforming the MINLP problem instance constraints into a mathematical structure that can be embedded directly onto the current D-Wave architecture. For testing and validation a D-Wave 2X device was used, as well as QxBranch’s QxLib software library, which includes a QC simulator based on simulated annealing. Our results indicate that it is mathematically possible to formulate the MINLP problem for AQO, but that currently available hardware is unable to solve problems of useful size. Classical general-purpose simulated annealing is currently able to solve larger problem sizes, but does not scale well and such methods would likely be outperformed in the future by improved AQO hardware with higher qubit connectivity and lower temperatures. If larger AQO devices are able to show improvements that trend in this direction, commercially viable solutions to the MINLP for particular applications could be implemented on hardware projected to be available in 5-10 years. Continued investigation into optimal AQO hardware architectures and novel methods for embedding MINLP problem constraints on to those architectures is needed to realize those commercial benefits.

Keywords: adiabatic quantum optimization, mixed integer nonlinear programming, quantum computing, NP-hard

Procedia PDF Downloads 526

Authors: Geetanjli Rani

Abstract:

The physical forms and network pattern of the city is expected to be enhanced with the advancement of technology. Reason being, the era of virtualisation and digital urban realm convergence with physical development. By means of comparative Spatial graphics and visuals of cities, the present paper attempts to revisit the very base of efficient physical forms and patterns to sync the emergence of virtual activities. Thus, the present approach to integrate spatial Smartness of Cities and Intelligent Transportation Systems is a brief assessment of smart forms and intelligent transportation network pattern to the dualism of physical and virtual urban activities. Finally, the research brings out that the grid iron pattern, radial, ring-radial, orbital etc. stands to be more efficient, effective and economical transit friendly for users, resource optimisation as well as compact urban and regional systems. Moreover, this paper concludes that the idea of flow and contiguity hidden in such smart forms and intelligent transportation network pattern suits to layering, deployment, installation and development of Intelligent Transportation Systems of Smart Cities such as infrastructure, facilities and services.

Keywords: smart form, smart infrastructure, intelligent transportation network pattern, physical and virtual integration

Procedia PDF Downloads 154

Authors: Devon Brown, Liu Chunmei

Abstract:

This paper will present an analysis as part of the researchers dissertation topic focusing on the intersection of affective and analytical directed acyclic graphs (DAGs) in the context of Decision Support Systems (DSS). The researcher’s work involves analyzing decision theory models like Affective and Bayesian Decision theory models and how they could be implemented under an Affective Computing Framework using Information Fusion and Human-Centered Design. Additionally, the researcher is beginning research on an Affective-Analytic Decision Framework (AADF) model for their dissertation research and are looking to merge logic and analytic models with empathetic insights into affective DAGs. Data-collection efforts begin Fall 2024 and in preparation for the efforts this paper looks to analyze previous research in this area and introduce the AADF framework and propose conceptual models for consideration. For this paper, the research emphasis is placed on analyzing Bayesian networks and Markov models which offer probabilistic techniques during uncertainty in decision-making. Ideally, including affect into analytic models will ensure algorithms can increase user trust with algorithms by including emotional states and the user’s experience with the goal of developing emotionally intelligent A.I. systems that can start to navigate the complex fabric of human emotion during decision-making.

Keywords: decision support systems, explainable AI, HCAI techniques, affective-analytical decision framework

Procedia PDF Downloads 20