Search results for: ocean models

Authors: Ghaleb Y. Abbasi, Israa Abu Rumman

Abstract:

This paper modeled sales history archived from 2012 to 2015 bulked in monthly bins for five products for a medical supply company in Jordan. The sales forecasts and extracted consistent patterns in the sales demand history from the Enterprise Resource Planning (ERP) system were used to predict future forecasting and generate sales demand forecasting using time series analysis statistical technique called Auto Regressive Integrated Moving Average (ARIMA). This was used to model and estimate realistic sales demand patterns and predict future forecasting to decide the best models for five products. Analysis revealed that the current replenishment system indicated inventory overstocking.

Keywords: ARIMA models, sales demand forecasting, time series, R code

Procedia PDF Downloads 385

Authors: Sean W. T. Bayly, Daniel Glover, Don Horrell, Simon Horrocks, Barnes Callum, Stuart Gibson, Mac Misuira

Abstract:

We assess suitability of recent, ∼7B parameter, instruction-tuned Language Models for Generative Named Entity Recognition (GNER). Alongside Retrieval Augmented Generation (RAG), and supported by task-specific prompting, our proposed Multi-Level Information Retrieval method achieves notable improvements over finetuned entity-level and sentence-level methods. We conclude that language models directed toward this task are highly capable when distinguishing between positive classes (precision). However, smaller models seem to struggle to find all entities (recall). Poorly defined classes such as ”Miscellaneous” exhibit substantial declines in performance, likely due to the ambiguity it introduces to the prompt. This is partially resolved through a self-verification method using engineered prompts containing knowledge of the stricter class definitions, particularly in areas where their boundaries are in danger of overlapping, such as the conflation between the location ”Britain” and the nationality ”British”. Finally, we explore correlations between model performance on the GNER task with performance on relevant academic benchmarks.

Keywords: generative named entity recognition, information retrieval, lightweight artificial intelligence, prompt engineering, personal information identification, retrieval augmented generation, self verification

Procedia PDF Downloads 49

Authors: R. Nandhini, Gaurab Mudbhari

Abstract:

Ranging from the field of health care to self-driving cars, machine learning and deep learning algorithms have revolutionized the field with the proper utilization of images and visual-oriented data. Segmentation, regression, classification, clustering, dimensionality reduction, etc., are some of the Machine Learning tasks that helped Machine Learning and Deep Learning models to become state-of-the-art models for the field where images are key datasets. Among these tasks, classification and clustering are essential but difficult because of the intricate and high-dimensional characteristics of image data. This finding examines and assesses advanced techniques in supervised classification and unsupervised clustering for image datasets, emphasizing the relative efficiency of Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), Deep Embedded Clustering (DEC), and self-supervised learning approaches. Due to the distinctive structural attributes present in images, conventional methods often fail to effectively capture spatial patterns, resulting in the development of models that utilize more advanced architectures and attention mechanisms. In image classification, we investigated both CNNs and ViTs. One of the most promising models, which is very much known for its ability to detect spatial hierarchies, is CNN, and it serves as a core model in our study. On the other hand, ViT is another model that also serves as a core model, reflecting a modern classification method that uses a self-attention mechanism which makes them more robust as this self-attention mechanism allows them to lean global dependencies in images without relying on convolutional layers. This paper evaluates the performance of these two architectures based on accuracy, precision, recall, and F1-score across different image datasets, analyzing their appropriateness for various categories of images. In the domain of clustering, we assess DEC, Variational Autoencoders (VAEs), and conventional clustering techniques like k-means, which are used on embeddings derived from CNN models. DEC, a prominent model in the field of clustering, has gained the attention of many ML engineers because of its ability to combine feature learning and clustering into a single framework and its main goal is to improve clustering quality through better feature representation. VAEs, on the other hand, are pretty well known for using latent embeddings for grouping similar images without requiring for prior label by utilizing the probabilistic clustering method.

Keywords: machine learning, deep learning, image classification, image clustering

Procedia PDF Downloads 12

Authors: Andrew J. Zacharias

Abstract:

The forest garden approach (FGA) was established by Trees for the Future (TREES) over the organization’s 30 years of agroforestry projects in Sub-Saharan Africa. This method transforms traditional agricultural systems into highly managed gardens that produce food and marketable products year-round. The effects of the FGA on food security, dietary diversity, and economic resilience have been measured closely, and TREES has begun to closely monitor the environmental impacts through the use of sensors mounted on unmanned aerial vehicles, commonly known as 'drones'. These drones collect thousands of pictures to create 3-D models in both the visible and the near-infrared wavelengths. Analysis of these models provides TREES with quantitative and qualitative evidence of improvements to the annual above-ground biomass and leaf area indices, as measured in-situ using NDVI calculations.

Keywords: agroforestry, biomass, drones, NDVI

Procedia PDF Downloads 157

Authors: Christopher Godwin Udomboso, Angela Unna Chukwu, Isaac Kwame Dontwi

Abstract:

In selecting a Statistical Neural Network model, the Network Information Criterion (NIC) has been observed to be sample biased, because it does not account for sample sizes. The selection of a model from a set of fitted candidate models requires objective data-driven criteria. In this paper, we derived and investigated the Adjusted Network Information Criterion (ANIC), based on Kullback’s symmetric divergence, which has been designed to be an asymptotically unbiased estimator of the expected Kullback-Leibler information of a fitted model. The analyses show that on a general note, the ANIC improves model selection in more sample sizes than does the NIC.

Keywords: statistical neural network, network information criterion, adjusted network, information criterion, transfer function

Procedia PDF Downloads 567

Authors: Asmae Tafraouti, Yasmina Layouni, Pascal Kleimann

Abstract:

The growing market for Microsystems (MST) and Micro-Electromechanical Systems (MEMS) is driving the research for alternative manufacturing techniques to microelectronics-based technologies, which are generally expensive and time-consuming. Hot-embossing and micro-injection modeling of thermoplastics appear to be industrially viable processes. However, both require the use of master models, usually made in hard materials such as steel. These master models cannot be fabricated using standard microelectronics processes. Thus, other micromachining processes are used, such as laser machining or micro-electrical discharge machining (µEDM). In this work, µEDM has been used. The principle of µEDM is based on the use of a thin cylindrical micro-tool that erodes the workpiece surface. The two electrodes are immersed in a dielectric with a distance of a few micrometers (gap). When an electrical voltage is applied between the two electrodes, electrical discharges are generated, which cause material machining. In order to produce master models with high resolution and smooth surfaces, it is necessary to well control the discharge mechanism. However, several problems are encountered, such as a random electrical discharge process, the fluctuation of the discharge energy, the electrodes' polarity inversion, and the wear of the micro-tool. The effect of different parameters, such as the applied voltage, the working capacitor, the micro-tool diameter, and the initial gap, has been studied. This analysis helps to improve the machining performances, such as the workpiece surface condition and the lateral crater's gap.

Keywords: craters, electrical discharges, micro-electrical discharge machining, microsystems

Procedia PDF Downloads 74

Authors: Bader Alkandari, Jamal Y. Madouh, Ahmad M. Alkandari, Anwar A. Alnaqi

Abstract:

A new formulation of fuzzy linear estimation problem is presented. It is formulated as a linear programming problem. The objective is to minimize the spread of the data points, taking into consideration the type of the membership function of the fuzzy parameters to satisfy the constraints on each measurement point and to insure that the original membership is included in the estimated membership. Different models are developed for a fuzzy triangular membership. The proposed models are applied to different examples from the area of fuzzy linear regression and finally to different examples for estimating the electrical load on a busbar. It had been found that the proposed technique is more suited for electrical load estimation, since the nature of the load is characterized by the uncertainty and vagueness.

Keywords: fuzzy regression, load estimation, fuzzy linear parameters, electrical load estimation

Procedia PDF Downloads 540

Authors: N. P. Bhandary, R. C. Tiwari, R. Yatabe

Abstract:

The research employs finite element method to evaluate the stability of roadside settlements slopes from Barpak to Yamagaon through Laprak village of Gorkha, Nepal after the 7.8Mw 2015 Barpak, Gorkha, Nepal earthquake. It includes three major villages of Gorkha, i.e., Barpak, Laprak and Yamagaun that were devastated by 2015 Gorkhas’ earthquake. The road head distance from the Barpak to Laprak and Laprak to Yamagaun are about 14 and 29km respectively. The epicentral distance of main shock of magnitude 7.8 and aftershock of magnitude 6.6 were respectively 7 and 11 kilometers (South-East) far from the Barpak village nearer to Laprak and Yamagaon. It is also believed that the epicenter of the main shock as said until now was not in the Barpak village, it was somewhere near to the Yamagaun village. The chaos that they had experienced during the earthquake in the Yamagaun was much more higher than the Barpak. In this context, we have carried out a detailed study to investigate the stability of Yamagaun settlements slope as a case study, where ground fissures, ground settlement, multiple cracks and toe failures are the most severe. In this regard, the stability issues of existing settlements and proposed road alignment, on the Yamagaon village slope are addressed, which is surrounded by many newly activated landslides. Looking at the importance of this issue, field survey is carried out to understand the behavior of ground fissures and multiple failure characteristics of the slopes. The results suggest that the Yamgaun slope in Profile 2-2, 3-3 and 4-4 are not safe enough for infrastructure development even in the normal soil slope conditions as per 2, 3 and 4 material models; however, the slope seems quite safe for at Profile 1-1 for all 4 material models. The result also indicates that the first three profiles are marginally safe for 2, 3 and 4 material models respectively. The Profile 4-4 is not safe enough for all 4 material models. Thus, Profile 4-4 needs a special care to make the slope stable.

Keywords: earthquake, finite element method, landslide, stability

Procedia PDF Downloads 348

Authors: Eduardo C. Guardia, Jose W. M. Lima, Afonso H. M. Santos

Abstract:

This paper presents a nonparametric method to obtain the hazard rate “Bathtub curve” for power system components. The model is a mixture of the three known phases of a component life, the decreasing failure rate (DFR), the constant failure rate (CFR) and the increasing failure rate (IFR) represented by three parametric Weibull models. The parameters are obtained from a simultaneous fitting process of the model to the Kernel nonparametric hazard rate curve. From the Weibull parameters and failure rate curves the useful lifetime and the characteristic lifetime were defined. To demonstrate the model the historic time-to-failure of distribution transformers were used as an example. The resulted “Bathtub curve” shows the failure rate for the equipment lifetime which can be applied in economic and replacement decision models.

Keywords: bathtub curve, failure analysis, lifetime estimation, parameter estimation, Weibull distribution

Procedia PDF Downloads 446

Authors: Aniket Kumar, Jacob Peedicayil

Abstract:

Major depressive disorder (MDD) is a common and important psychiatric disorder. Several clinical features of MDD suggest an epigenetic basis for its pathogenesis. Since epigenetics (heritable changes in gene expression not involving changes in DNA sequence) may underlie the pathogenesis of MDD, epigenetic drugs such as DNA methyltransferase inhibitors (DNMTi) and histone deactylase inhibitors (HDACi) may be useful for treating MDD. The available literature indexed in Pubmed on preclinical drug trials of epigenetic drugs for the treatment of MDD was investigated. The search terms we used were ‘depression’ or ‘depressive’ and ‘HDACi’ or ‘DNMTi’. Among epigenetic drugs, it was found that there were 3 preclinical trials using HDACi and 3 using DNMTi for the treatment of MDD. All the trials were conducted on rodents (mice or rats). The animal models of depression that were used were: learned helplessness-induced animal model, forced swim test, open field test, and the tail suspension test. One study used a genetic rat model of depression (the Flinders Sensitive Line). The HDACi that were tested were: sodium butyrate, compound 60 (Cpd-60), and valproic acid. The DNMTi that were tested were: 5-azacytidine and decitabine. Among the three preclinical trials using HDACi, all showed an antidepressant effect in animal models of depression. Among the 3 preclinical trials using DNMTi also, all showed an antidepressant effect in animal models of depression. Thus, epigenetic drugs, namely, HDACi and DNMTi, may prove to be useful in the treatment of MDD and merit further investigation for the treatment of this disorder.

Keywords: DNA methylation, drug discovery, epigenetics, major depressive disorder

Procedia PDF Downloads 188

Authors: Yin Zhang, Kai Qiao, Xiyang Zhi, Jinnan Gong, Jianming Hu

Abstract:

In order to achieve effective detection of aerial targets over long distances from space-based platforms, the mechanism of interaction between the radiation characteristics of the aerial targets and the complex scene environment including the sunlight conditions, underlying surfaces and the atmosphere are analyzed. A large simulated database of space-based radiance images is constructed considering several typical aerial targets, target working modes (flight velocity and altitude), illumination and observation angles, background types (cloud, ocean, and urban areas) and sensor spectrums ranging from visible to thermal infrared. The target detectability is characterized by the signal-to-clutter ratio (SCR) extracted from the images. The influence laws of the target detectability are discussed under different detection bands and instantaneous fields of view (IFOV). Furthermore, the optimal center wavelengths and widths of the detection bands are suggested, and the minimum IFOV requirements are proposed. The research can provide theoretical support and scientific guidance for the design of space-based detection systems and on-board information processing algorithms.

Keywords: space-based detection, aerial targets, detectability analysis, scene environment

Procedia PDF Downloads 144

Authors: Joao Fialho

Abstract:

The mathematical modelling of idiopathic scoliosis has been studied throughout the years. The models presented on those papers are based on the orthotic stabilization of the idiopathic scoliosis, which are based on a transversal force being applied to the human spine on a continuous form. When considering the ATT (Antalgic-Trak Technology) device, the existent models cannot be used, as the type of forces applied are no longer transversal nor applied in a continuous manner. In this device, vertical traction is applied. In this study we propose to model the idiopathic scoliosis, using the ATT (Antalgic-Trak Technology) device, and with the parameters obtained from the mathematical modeling, set up a case-by-case individualized therapy plan, for each patient.

Keywords: idiopathic scoliosis, mathematical modelling, human spine, Antalgic-Trak technology

Procedia PDF Downloads 269

Authors: Mohamed A. G. H. Abdalsadig, Amir Nourian, G. G. Nasr, M. Babaie

Abstract:

Gas lift is one of the most common forms of artificial lift, particularly for offshore wells because of its relative down hole simplicity, flexibility, reliability, and ability to operate over a large range of rates and occupy very little space at the well head. Presently, petroleum industry is investing in exploration and development fields in offshore locations where oil and gas wells are being drilled thousands of feet below the ocean in high pressure and temperature conditions. Therefore, gas-lifted oil wells are capable of failure through gas lift valves which are considered as the heart of the gas lift system for controlling the amount of the gas inside the tubing string. The gas injection rate through gas lift valve must be controlled to be sufficient to obtain and maintain critical flow, also, gas lift valves must be designed not only to allow gas passage through it and prevent oil passage, but also for gas injection into wells to be started and stopped when needed. In this paper, smart gas lift valve has been used to investigate the effect of the valve port size, depth of injection and vertical lift performance on well productivity; all these aspects have been investigated using PROSPER simulator program coupled with experimental data. The results show that by using smart gas lift valve, the gas injection rate can be controlled which leads to improved flow performance.

Keywords: Effect of gas lift valve port size, effect water cut, vertical flow performance

Procedia PDF Downloads 292

Authors: Shahram Jamali, Samira Hamed

Abstract:

One of the open issues in Random Early Detection (RED) algorithm is how to set its parameters to reach high performance for the dynamic conditions of the network. Although original RED uses fixed values for its parameters, this paper follows a model-based approach to upgrade performance of the RED algorithm. It models the routers queue behavior by using the Markov model and uses this model to predict future conditions of the queue. This prediction helps the proposed algorithm to make some tunings over RED's parameters and provide efficiency and better performance. Widespread packet level simulations confirm that the proposed algorithm, called Markov-RED, outperforms RED and FARED in terms of queue stability, bottleneck utilization and dropped packets count.

Keywords: active queue management, RED, Markov model, random early detection algorithm

Procedia PDF Downloads 539

Authors: Andrey Khalov

Abstract:

The rapid proliferation of unstructured data in IT infrastructure management demands innovative approaches for extracting actionable knowledge. This paper presents a framework for ontology-based knowledge extraction that combines relational graph neural networks (R-GNN) with large language models (LLMs). The proposed method leverages the DOLCE framework as the foundational ontology, extending it with concepts from ITSMO for domain-specific applications in IT service management and outsourcing. A key component of this research is the use of transformer-based models, such as DeBERTa-v3-large, for automatic entity and relationship extraction from unstructured texts. Furthermore, the paper explores how transfer learning techniques can be applied to fine-tune large language models (LLaMA) for using to generate synthetic datasets to improve precision in BERT-based entity recognition and ontology alignment. The resulting IT Ontology (ITO) serves as a comprehensive knowledge base that integrates domain-specific insights from ITIL processes, enabling more efficient decision-making. Experimental results demonstrate significant improvements in knowledge extraction and relationship mapping, offering a cutting-edge solution for enhancing cognitive computing in IT service environments.

Keywords: ontology expansion, synthetic dataset, transformer fine-tuning, concept extraction, DOLCE, BERT, taxonomy, LLM, NER

Procedia PDF Downloads 14

Authors: Sofia Stoica

Abstract:

In this paper, we use historical data to predict the stock price of a tech company. To this end, we use a dataset consisting of the stock prices in the past five years of ten major tech companies – Adobe, Amazon, Apple, Facebook, Google, Microsoft, Netflix, Oracle, Salesforce, and Tesla. We experimented with a variety of models– a linear regressor model, K nearest Neighbors (KNN), a sequential neural network – and algorithms - Multiplicative Weight Update, and AdaBoost. We found that the sequential neural network performed the best, with a testing error of 0.18%. Interestingly, the linear model performed the second best with a testing error of 0.73%. These results show that using historical data is enough to obtain high accuracies, and a simple algorithm like linear regression has a performance similar to more sophisticated models while taking less time and resources to implement.

Keywords: finance, machine learning, opening price, stock market

Procedia PDF Downloads 191

Authors: Farzam Fatolazadeh, Kalifa Goita, Mehdi Eshagh, Shusen Wang

Abstract:

The Gravity Recovery and Climate Experiment (GRACE) is a satellite mission with twin satellites for the precise determination of spatial and temporal variations in the Earth’s gravity field. The products of this mission are monthly global gravity models containing the spherical harmonic coefficients and their errors. These GRACE models can be used for estimating terrestrial water storage (TWS) variations across the globe at large scales, thereby offering an opportunity for surface and groundwater storage (GWS) assessments. Yet, the ability of GRACE to monitor changes at smaller scales is too limited for local water management authorities. This is largely due to the low spatial and temporal resolutions of its models (~200,000 km2 and one month, respectively). High-resolution GRACE data products would substantially enrich the information that is needed by local-scale decision-makers while offering the data for the regions that lack adequate in situ monitoring networks, including northern parts of Canada. Such products could eventually be obtained through downscaling. In this study, we extended the spectral combination theory to simultaneously downscale spatiotemporally the 3o spatial coarse resolution of GRACE to 0.25o degrees resolution and monthly coarse resolution to daily resolution. This method combines the monthly gravity field solution of GRACE and daily hydrological model products in the form of both low and high-frequency signals to produce high spatiotemporal resolution TWSA and GWSA products. The main contribution and originality of this study are to comprehensively and simultaneously consider GRACE and hydrological variables and their uncertainties to form the estimator in the spectral domain. Therefore, it is predicted that we reach downscale products with an acceptable accuracy.

Keywords: GRACE satellite, groundwater storage, spectral combination, terrestrial water storage

Procedia PDF Downloads 83

Authors: Taghreed Alghamdi, Wendy Hall, David Millard

Abstract:

Massive Open Online Courses (MOOCs) have recently gained widespread interest in the academic world, starting a wide range of discussion of a number of issues. One of these issues, using MOOCs in teaching and learning in the higher education by integrating MOOCs’ contents with traditional face-to-face activities in blended learning format, is called blended MOOCs (bMOOCs) and is intended not to replace traditional learning but to enhance students learning. Most research on MOOCs has focused on students’ perception and institutional threats whereas there is a lack of published research on academics’ experiences and practices. Thus, the first aim of the study is to develop a classification of blended MOOCs models by conducting a systematic literature review, classifying 19 different case studies, and identifying the broad types of bMOOCs models namely: Supplementary Model and Integrated Model. Thus, the analyses phase will emphasize on these different types of bMOOCs models in terms of adopting MOOCs by lecturers. The second aim of the study is to improve the understanding of lecturers’ acceptance of bMOOCs by investigate the factors that influence academics’ acceptance of using MOOCs in traditional learning by distributing an online survey to lecturers who participate in MOOCs platforms. These factors can help institutions to encourage their lecturers to integrate MOOCs with their traditional courses in universities.

Keywords: acceptance, blended learning, blended MOOCs, higher education, lecturers, MOOCs, professors

Procedia PDF Downloads 131

Authors: Aasheesh Raturi, Vimal Kothiyal, P. D. Semalty

Abstract:

We studied mechanical properties of Eucalyptus tereticornis using FT-NIR spectroscopy. Firstly, spectra were pre-processed to eliminate useless information. Then, prediction model was constructed by partial least squares regression. To study the influence of pre-processing on prediction of mechanical properties for NIR analysis of wood samples, we applied various pretreatment methods like straight line subtraction, constant offset elimination, vector-normalization, min-max normalization, multiple scattering. Correction, first derivative, second derivatives and their combination with other treatment such as First derivative + straight line subtraction, First derivative+ vector normalization and First derivative+ multiplicative scattering correction. The data processing methods in combination of preprocessing with different NIR regions, RMSECV, RMSEP and optimum factors/rank were obtained by optimization process of model development. More than 350 combinations were obtained during optimization process. More than one pre-processing method gave good calibration/cross-validation and prediction/test models, but only the best calibration/cross-validation and prediction/test models are reported here. The results show that one can safely use NIR region between 4000 to 7500 cm-1 with straight line subtraction, constant offset elimination, first derivative and second derivative preprocessing method which were found to be most appropriate for models development.

Keywords: FT-NIR, mechanical properties, pre-processing, PLS

Procedia PDF Downloads 362

Authors: Rimon Rafiah

Abstract:

This paper will present a combination of two seemingly unrelated models, which are the one for estimating economic development impacts as a result of transportation investment and the other for increasing CAV penetration in order to reduce congestion. Measuring economic development impacts resulting from transportation investments is becoming more recognized around the world. Examples include the UK’s Wider Economic Benefits (WEB) model, Economic Impact Assessments in the USA, various input-output models, and additional models around the world. The economic impact model is based on WEB and is based on the following premise: investments in transportation will reduce the cost of personal travel, enabling firms to be more competitive, creating additional throughput (the same road allows more people to travel), and reducing the cost of travel of workers to a new workplace. This reduction in travel costs was estimated in out-of-pocket terms in a given localized area and was then translated into additional employment based on regional labor supply elasticity. This additional employment was conservatively assumed to be at minimum wage levels, translated into GDP terms, and from there into direct taxation (i.e., an increase in tax taken by the government). The CAV model is based on economic principles such as CAV usage, supply, and demand. Usage of CAVs can increase capacity using a variety of means – increased automation (known as Level I thru Level IV) and also by increased penetration and usage, which has been predicted to go up to 50% by 2030 according to several forecasts, with possible full conversion by 2045-2050. Several countries have passed policies and/or legislation on sales of gasoline-powered vehicles (none) starting in 2030 and later. Supply was measured via increased capacity on given infrastructure as a function of both CAV penetration and implemented technologies. The CAV model, as implemented in the USA, has shown significant savings in travel time and also in vehicle operating costs, which can be translated into economic development impacts in terms of job creation, GDP growth and salaries as well. The models have policy implications as well and can be adapted for use in Japan as well.

Keywords: CAV, economic development, WEB, transport economics

Procedia PDF Downloads 74

Authors: Dewi Selvia Fardhyanti, Astrilia Damayanti

Abstract:

Coal tar is a liquid by-product of the process of coal gasification and carbonation, also in some industries such as steel, power plant, cement, and others. This liquid oil mixture contains various kinds of useful compounds such as aromatic compounds and phenolic compounds. These compounds are widely used as raw material for insecticides, dyes, medicines, perfumes, coloring matters, and many others. This research investigates thermodynamic modelling of liquid-liquid equilibria (LLE) in solvent extraction of o-Cresol from the coal tar. The equilibria are modeled by ternary components of Wohl, Van Laar, and Three-Suffix Margules models. The values of the parameters involved are obtained by curve-fitting to the experimental data. Based on the comparison between calculated and experimental data, it turns out that among the three models studied, the Three-Suffix Margules seems to be the best to predict the LLE of o-Cresol for those system.

Keywords: coal tar, o-Cresol, Wohl, Van Laar, three-suffix margules

Procedia PDF Downloads 277

Authors: Parsa Mahdavi, M. Amin Hariri-Ardebili

Abstract:

The development of accurate machine learning and deep learning models traditionally demands hands-on expertise and a solid background to fine-tune hyperparameters. With the continuous expansion of datasets in various scientific and engineering domains, researchers increasingly turn to machine learning methods to unveil hidden insights that may elude classic regression techniques. This surge in adoption raises concerns about the adequacy of the resultant meta-models and, consequently, the interpretation of the findings. In response to these challenges, automated machine learning (AutoML) emerges as a promising solution, aiming to construct machine learning models with minimal intervention or guidance from human experts. AutoML encompasses crucial stages such as data preparation, feature engineering, hyperparameter optimization, and neural architecture search. This paper provides a comprehensive overview of the principles underpinning AutoML, surveying several widely-used AutoML platforms. Additionally, the paper offers a glimpse into the application of AutoML on various engineering datasets. By comparing these results with those obtained through classical machine learning methods, the paper quantifies the uncertainties inherent in the application of a single ML model versus the holistic approach provided by AutoML. These examples showcase the efficacy of AutoML in extracting meaningful patterns and insights, emphasizing its potential to revolutionize the way we approach and analyze complex datasets.

Keywords: automated machine learning, uncertainty, engineering dataset, regression

Procedia PDF Downloads 61

Authors: Ramazan I. Kadiev, Arcady Ponosov

Abstract:

Mathematical models of gene regulatory networks can in many cases be described by ordinary differential equations with switching nonlinearities, where the initial value problem is ill-posed. Several regularization methods are known in the case of deterministic networks, but the presence of stochastic noise leads to several technical difficulties. In the presentation, it is proposed to apply the methods of the stochastic singular perturbation theory going back to Yu. Kabanov and Yu. Pergamentshchikov. This approach is used to regularize the above ill-posed problem, which, e.g., makes it possible to design stable numerical schemes. Several examples are provided in the presentation, which support the efficiency of the suggested analysis. The method can also be of interest in other fields of biomathematics, where differential equations contain switchings, e.g., in neural field models.

Keywords: ill-posed problems, singular perturbation analysis, stochastic differential equations, switching nonlinearities

Procedia PDF Downloads 194

Authors: Ishmael Tingbani

Abstract:

This paper adopts a contingency theory approach to investigate the relationship between working capital management and profitability using data of 225 listed British firms on the London Stock Exchange for the period 2001-2011. The paper employs a panel data analysis on a series of interactive models to estimate this relationship. The findings of the study confirm the relevance of the contingency theory. Evidence from the study suggests that the impact of working capital management on profitability varies and is constrained by organizational contingencies (environment, resources, and management factors) of the firm. These findings have implications for a more balanced and nuanced view of working capital management policy for policy-makers.

Keywords: working capital management, profitability, contingency theory approach, interactive models

Procedia PDF Downloads 347

Authors: Asmae Tafraouti, Yasmina Layouni, Pascal Kleimann

Abstract:

The growing market for Microsystems (MST) and Micro-Electromechanical Systems (MEMS) is driving the research for alternative manufacturing techniques to microelectronics-based technologies, which are generally expensive and time-consuming. Hot-embossing and micro-injection modeling of thermoplastics appear to be industrially viable processes. However, both require the use of master models, usually made in hard materials such as steel. These master models cannot be fabricated using standard microelectronics processes. Thus, other micromachining processes are used, as laser machining or micro-electrical discharge machining (µEDM). In this work, µEDM has been used. The principle of µEDM is based on the use of a thin cylindrical micro-tool that erodes the workpiece surface. The two electrodes are immersed in a dielectric with a distance of a few micrometers (gap). When an electrical voltage is applied between the two electrodes, electrical discharges are generated, which cause material machining. In order to produce master models with high resolution and smooth surfaces, it is necessary to well control the discharge mechanism. However, several problems are encountered, such as a random electrical discharge process, the fluctuation of the discharge energy, the electrodes' polarity inversion, and the wear of the micro-tool. The effect of different parameters, such as the applied voltage, the working capacitor, the micro-tool diameter, the initial gap, has been studied. This analysis helps to improve the machining performances, such: the workpiece surface condition and the lateral crater's gap.

Keywords: craters, electrical discharges, micro-electrical discharge machining (µEDM), microsystems

Procedia PDF Downloads 96

Authors: Endrik Mifta Shaiful, Firman Riyudha

Abstract:

Acquired Immunodeficiency Syndrome (AIDS) is one of the world's deadliest diseases caused by the Human Immunodeficiency Virus (HIV) that infects white blood cells and cause a decline in the immune system. AIDS quickly became a world epidemic disease that affects almost all countries. Therefore, mathematical modeling approach to the spread of HIV and AIDS is needed to anticipate the spread of HIV and AIDS which are widespread. The purpose of this study is to determine the parameter estimation on mathematical models of HIV transmission and AIDS using cumulative data of people with HIV and AIDS each year in Indonesia. In this model, there are parameters of r ∈ [0,1) which is the effectiveness of the treatment in patients with HIV. If the value of r is close to 1, the number of people with HIV and AIDS will decline toward zero. The estimation results indicate when the value of r is close to unity, there will be a significant decline in HIV patients, whereas in AIDS patients constantly decreases towards zero.

Keywords: HIV, AIDS, parameter estimation, mathematical models

Procedia PDF Downloads 251

Authors: V. Mantey, N. Findlay, I. Maddox

Abstract:

The quality and quantity of global satellite data have been increasing exponentially in recent years as spaceborne systems become more affordable and the sensors themselves become more sophisticated. This is a valuable resource for many applications, including disaster management and relief. However, while more information can be valuable, the volume of data available is impossible to manually examine. Therefore, the question becomes how to extract as much information as possible from the data with limited manpower. Buildings are a key feature of interest in satellite imagery with applications including telecommunications, population models, and disaster relief. Machine learning tools are fast becoming one of the key resources to solve this problem, and models have been developed to detect buildings in optical satellite imagery. However, by and large, most models focus on affluent regions where buildings are generally larger and constructed further apart. This work is focused on the more difficult problem of detection in populated regions. The primary challenge with detecting small buildings in densely populated regions is both the spatial and spectral resolution of the optical sensor. Densely packed buildings with similar construction materials will be difficult to separate due to a similarity in color and because the physical separation between structures is either non-existent or smaller than the spatial resolution. This study finds that training models until they are overfitting the input sample can perform better in these areas than a more robust, generalized model. An overfitted model takes less time to fine-tune from a generalized pre-trained model and requires fewer input data. The model developed for this study has also been fine-tuned using existing, open-source, building vector datasets. This is particularly valuable in the context of disaster relief, where information is required in a very short time span. Leveraging existing datasets means that little to no manpower or time is required to collect data in the region of interest. The training period itself is also shorter for smaller datasets. Requiring less data means that only a few quality areas are necessary, and so any weaknesses or underpopulated regions in the data can be skipped over in favor of areas with higher quality vectors. In this study, a landcover classification model was developed in conjunction with the building detection tool to provide a secondary source to quality check the detected buildings. This has greatly reduced the false positive rate. The proposed methodologies have been implemented and integrated into a configurable production environment and have been employed for a number of large-scale commercial projects, including continent-wide DEM production, where the extracted building footprints are being used to enhance digital elevation models. Overfitted machine learning models are often considered too specific to have any predictive capacity. However, this study demonstrates that, in cases where input data is scarce, overfitted models can be judiciously applied to solve time-sensitive problems.

Keywords: building detection, disaster relief, mask-RCNN, satellite mapping

Procedia PDF Downloads 169

Authors: Malinda Gayan Sirisena

Abstract:

The aim of this research is to evaluate the effectiveness of software quality assurance approaches of Sri Lankan offshore software development organizations, and to propose a framework which could be used across all offshore software development organizations. An empirical study was conducted using derived framework from popular software quality evaluation models. The research instrument employed was a questionnaire survey among thirty seven Sri Lankan registered offshore software development organizations. The findings demonstrate a positive view of Effectiveness of Software Quality Assurance – the stronger predictors of Stability, Installability, Correctness, Testability and Changeability. The present study’s recommendations indicate a need for much emphasis on software quality assurance for the Sri Lankan offshore software development organizations.

Keywords: software quality assurance (SQA), offshore software development, quality assurance evaluation models, effectiveness of quality assurance

Procedia PDF Downloads 423

Authors: Charles Horace Ampong

Abstract:

Non-linear models have been found to be useful in modeling the elasticity (measure of degree of responsiveness) of a dependent variable with respect to a set of independent variables ceteris paribus. This constant elasticity principle was applied to the dependent variable (Compressive Strength of Concrete in MPa) which was found to be non-linearly related to the independent variable (Water-Cement ratio in kg/m3) for given Ages of Concrete in days (3, 7, 28) at different levels of admixtures Superplasticizer (in kg/m3), Blast Furnace Slag (in kg/m3) and Fly Ash (in kg/m3). The levels of the admixtures were categorized as: S1=Some Plasticizer added & S0=No Plasticizer added; B1=some Blast Furnace Slag added & B0=No Blast Furnace Slag added; F1=Some Fly Ash added & F0=No Fly Ash added. The number of observations (samples) used for the research was one-hundred and thirty-two (132) in all. For Superplasticizer, it was found that Compressive Strength of Concrete was more elastic with regards to Water-Cement ratio at S1 level than at S0 level for the given ages of concrete 3, 7and 28 days. For Blast Furnace Slag, Compressive Strength with regards to Water-Cement ratio was more elastic at B0 level than at B1 level for concrete ages 3, 7 and 28 days. For Fly Ash, Compressive Strength with regards to Water-Cement ratio was more elastic at B0 level than at B1 level for Ages 3, 7 and 28 days. The research also tested for different combinations of the levels of Superplasticizer, Blast Furnace Slag and Fly Ash. It was found that Compressive Strength elasticity with regards to Water-Cement ratio was lowest (Elasticity=-1.746) with a combination of S0, B0 and F0 for concrete age of 3 days. This was followed by Elasticity of -1.611 with a combination of S0, B0 and F0 for a concrete of age 7 days. Next, the highest was an Elasticity of -1.414 with combination of S0, B0 and F0 for a concrete age of 28 days. Based on preceding outcomes, three (3) non-linear model equations for predicting the output elasticity of Compressive Strength of Concrete (in %) or the value of Compressive Strength of Concrete (in MPa) with regards to Water to Cement was formulated. The model equations were based on the three different ages of concrete namely 3, 7 and 28 days under investigation. The three models showed that higher elasticity translates into higher compressive strength. And the models revealed a trend of increasing concrete strength from 3 to 28 days for a given amount of water to cement ratio. Using the models, an increasing modulus of elasticity from 3 to 28 days was deduced.

Keywords: concrete, compressive strength, elasticity, water-cement

Procedia PDF Downloads 293

Authors: Dingyang Hu, Dan Liu

Abstract:

DNN (Deep Neural Network) deep learning models are widely used in classification, prediction, and other task scenarios. To address the difficulties of generic adversarial perturbation generation for deep learning models under black-box conditions, a generic adversarial ingestion generation method based on a saliency map (CJsp) is proposed to obtain salient image regions by counting the factors that influence the input features of an image on the output results. This method can be understood as a saliency map attack algorithm to obtain false classification results by reducing the weights of salient feature points. Experiments also demonstrate that this method can obtain a high success rate of migration attacks and is a batch adversarial sample generation method.

Keywords: adversarial sample, gradient, probability, black box

Procedia PDF Downloads 104