Search results for: forecasting accuracy

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 176

Authors: Loai AbdAllah, Mahmoud Kaiyal

Abstract:

Missing values in real-world datasets are a common problem. Many algorithms were developed to deal with this problem, most of them replace the missing values with a fixed value that was computed based on the observed values. In our work, we used a distance function based on Bhattacharyya distance to measure the distance between objects with missing values. Bhattacharyya distance, which measures the similarity of two probability distributions. The proposed distance distinguishes between known and unknown values. Where the distance between two known values is the Mahalanobis distance. When, on the other hand, one of them is missing the distance is computed based on the distribution of the known values, for the coordinate that contains the missing value. This method was integrated with Wikaya, a digital health company developing a platform that helps to improve prevention of chronic diseases such as diabetes and cancer. In order for Wikaya’s recommendation system to work distance between users need to be measured. Since there are missing values in the collected data, there is a need to develop a distance function distances between incomplete users profiles. To evaluate the accuracy of the proposed distance function in reflecting the actual similarity between different objects, when some of them contain missing values, we integrated it within the framework of k nearest neighbors (kNN) classifier, since its computation is based only on the similarity between objects. To validate this, we ran the algorithm over diabetes and breast cancer datasets, standard benchmark datasets from the UCI repository. Our experiments show that kNN classifier using our proposed distance function outperforms the kNN using other existing methods.

Keywords: missing values, incomplete data, distance, incomplete diabetes data

Procedia PDF Downloads 205

Authors: Emmanuel Osaze Oshoiribhor, Adetokunbo MacGregor John-Otumu

Abstract:

In recent years, there has been a desire to forecast student academic achievement prior to graduation. This is to help them improve their grades, particularly for individuals with poor performance. The goal of this study is to employ supervised learning techniques to construct a predictive model for student academic achievement. Many academics have already constructed models that predict student academic achievement based on factors such as smoking, demography, culture, social media, parent educational background, parent finances, and family background, to name a few. This feature and the model employed may not have correctly classified the students in terms of their academic performance. This model is built using a logistic regression classifier with basic features such as the previous semester's course score, attendance to class, class participation, and the total number of course materials or resources the student is able to cover per semester as a prerequisite to predict if the student will perform well in future on related courses. The model outperformed other classifiers such as Naive bayes, Support vector machine (SVM), Decision Tree, Random forest, and Adaboost, returning a 96.7% accuracy. This model is available as a desktop application, allowing both instructors and students to benefit from user-friendly interfaces for predicting student academic achievement. As a result, it is recommended that both students and professors use this tool to better forecast outcomes.

Keywords: artificial intelligence, ML, logistic regression, performance, prediction

Procedia PDF Downloads 84

Authors: S. Jain, R. Savalia, V. Saini

Abstract:

A simple, accurate, precise, sensitive and specific RP-HPLC method was developed and validated for simultaneous estimation of Metoprolol Succinate and Olmesartan Medoxomil in bulk and tablet dosage form. The RP-HPLC method has shown adequate separation for Metoprolol Succinate and Olmesartan Medoxomil from its degradation products. The separation was achieved on a Phenomenex luna ODS C18 (250mm X 4.6mm i.d., 5μm particle size) with an isocratic mixture of acetonitrile: 50mM phosphate buffer pH 4.0 adjusted with glacial acetic acid in the ratio of 55:45 v/v. The mobile phase at a flow rate of 1.0ml/min, Injection volume 20μl and wavelength of detection was kept at 225nm. The retention time for Metoprolol Succinate and Olmesartan Medoxomil was 2.451±0.1min and 6.167±0.1min, respectively. The linearity of the proposed method was investigated in the range of 5-50μg/ml and 2-20μg/ml for Metoprolol Succinate and Olmesartan Medoxomil, respectively. Correlation coefficient was 0.999 and 0.9996 for Metoprolol Succinate and Olmesartan Medoxomil, respectively. The limit of detection was 0.2847μg/ml and 0.1251μg/ml for Metoprolol Succinate and Olmesartan Medoxomil, respectively and the limit of quantification was 0.8630μg/ml and 0.3793μg/ml for Metoprolol and Olmesartan, respectively. Proposed methods were validated as per ICH guidelines for linearity, accuracy, precision, specificity and robustness for estimation of Metoprolol Succinate and Olmesartan Medoxomil in commercially available tablet dosage form and results were found to be satisfactory. Thus the developed and validated stability indicating method can be used successfully for marketed formulations.

Keywords: metoprolol succinate, olmesartan medoxomil, RP-HPLC method, validation, ICH

Procedia PDF Downloads 298

Authors: Nahla H. Alaswadko, Rayya A. Hassan, Bayar N. Mohammed

Abstract:

Empirical deterministic models have been developed to predict roughness progression of heavy duty spray sealed pavements for a dataset representing rural arterial roads. The dataset provides a good representation of the relevant network and covers a wide range of operating and environmental conditions. A sample with a large size of historical time series data for many pavement sections has been collected and prepared for use in multilevel regression analysis. The modelling parameters include road roughness as performance parameter and traffic loading, time, initial pavement strength, reactivity level of subgrade soil, climate condition, and condition of drainage system as predictor parameters. The purpose of this paper is to report the approaches adopted for models development and validation. The study presents multilevel models that can account for the correlation among time series data of the same section and to capture the effect of unobserved variables. Study results show that the models fit the data very well. The contribution and significance of relevant influencing factors in predicting roughness progression are presented and explained. The paper concludes that the analysis approach used for developing the models confirmed their accuracy and reliability by well-fitting to the validation data.

Keywords: roughness progression, empirical model, pavement performance, heavy duty pavement

Procedia PDF Downloads 159

Authors: Juhee Lee, Yongjun Lee

Abstract:

In designing a low-energy-consuming buildings, the heat transfer through a large glass or wall becomes critical. Multiple layers of the window glasses and walls are employed for the high insulation. The gravity driven air flow between window glasses or wall layers is a natural heat convection phenomenon being a key of the heat transfer. For the first step of the natural heat transfer analysis, in this study the development and application of a finite volume method for the numerical computation of viscous incompressible flows is presented. It will become a part of the natural convection analysis with high-order scheme, multi-grid method, and dual-time step in the future. A finite volume method based on a fully-implicit second-order is used to discretize and solve the fluid flow on unstructured grids composed of arbitrary-shaped cells. The integrations of the governing equation are discretised in the finite volume manner using a collocated arrangement of variables. The convergence of the SIMPLE segregated algorithm for the solution of the coupled nonlinear algebraic equations is accelerated by using a sparse matrix solver such as BiCGSTAB. The method used in the present study is verified by applying it to some flows for which either the numerical solution is known or the solution can be obtained using another numerical technique available in the other researches. The accuracy of the method is assessed through the grid refinement.

Keywords: finite volume method, fluid flow, laminar flow, unstructured grid

Procedia PDF Downloads 272

Authors: Raymond Xu, Cindy Jingru Wang

Abstract:

Social media has become a globally connected and influencing platform. Social media data, such as tweets, can help predict the spread of pandemics and provide individuals and healthcare providers early warnings. Public psychological reactions and opinions can be efficiently monitored by AI models on the progression of dominant topics on Twitter. However, statistics show that as the coronavirus spreads, so does an infodemic of misinformation due to pandemic-related factors such as unemployment and lockdowns. Social media algorithms are often biased toward outrage by promoting content that people have an emotional reaction to and are likely to engage with. This can influence users’ attitudes and cause confusion. Therefore, social media is a double-edged sword. Combating fake news and biased content has become one of the essential tasks. This research analyzes the variety of methods used for fake news detection covering random forest, logistic regression, support vector machines, decision tree, naive Bayes, BoW, TF-IDF, LDA, CNN, RNN, LSTM, DeepFake, and hierarchical attention network. The performance of each method is analyzed. Based on these models’ achievements and limitations, a multi-dimensional AI framework is proposed to achieve higher accuracy in infodemic detection, especially pandemic-related news. The model is trained on contextual content, images, and news metadata.

Keywords: artificial intelligence, fake news detection, infodemic detection, image recognition, sentiment analysis

Procedia PDF Downloads 228

Authors: Bruce X. B. Yu, Yan Liu, Keith C. C. Chan

Abstract:

We propose to record Activities of Daily Living (ADLs) of elderly people using a vision-based system so as to provide better assistive and personalization technologies. Current ADL-related research is based on data collected with help from non-elderly subjects in laboratory environments and the activities performed are predetermined for the sole purpose of data collection. To obtain more realistic datasets for the application, we recorded ADLs for the elderly with data collected from real-world environment involving real elderly subjects. Motivated by the need to collect data for more effective research related to elderly care, we chose to collect data in the room of an elderly person. Specifically, we installed Kinect, a vision-based sensor on the ceiling, to capture the activities that the elderly subject performs in the morning every day. Based on the data, we identified 12 morning activities that the elderly person performs daily. To recognize these activities, we created a HARELCARE framework to investigate into the effectiveness of existing Human Activity Recognition (HAR) algorithms and propose the use of a transfer learning algorithm for HAR. We compared the performance, in terms of accuracy, and training progress. Although the collected dataset is relatively small, the proposed algorithm has a good potential to be applied to all daily routine activities for healthcare purposes such as evidence-based diagnosis and treatment.

Keywords: daily activity recognition, healthcare, IoT sensors, transfer learning

Procedia PDF Downloads 121

Authors: Hicham Gourgue, Ahmed Aharoune, Ahmed Ihlal

Abstract:

We present in this work our model of road traffic emissions (line sources) and dispersion of these emissions, named DISPOLSPEM (Dispersion of Poly Sources and Pollutants Emission Model). In its emission part, this model was designed to keep the consistent bottom-up and top-down approaches. It also allows to generate emission inventories from reduced input parameters being adapted to existing conditions in Morocco and in the other developing countries. While several simplifications are made, all the performance of the model results are kept. A further important advantage of the model is that it allows the uncertainty calculation and emission rate uncertainty according to each of the input parameters. In the dispersion part of the model, an improved line source model has been developed, implemented and tested against a reference solution. It provides improvement in accuracy over previous formulas of line source Gaussian plume model, without being too demanding in terms of computational resources. In the case study presented here, the biggest errors were associated with the ends of line source sections; these errors will be canceled by adjacent sections of line sources during the simulation of a road network. In cases where the wind is parallel to the source line, the use of the combination discretized source and analytical line source formulas minimizes remarkably the error. Because this combination is applied only for a small number of wind directions, it should not excessively increase the calculation time.

Keywords: air pollution, dispersion, emissions, line sources, road traffic, urban transport

Procedia PDF Downloads 429

Authors: Spandan Sahu, Amiya Kumar Pati, Kishanjit Kumar Khatua

Abstract:

River is the main source of water. It is a form of natural open channel which gives rise to many complex phenomenon of sciences that needs to be tackled such as the critical flow conditions, boundary shear stress and depth averaged velocity. The development of society more or less solely depends upon the flow of rivers. The rivers are major sources of many sediments and specific ingredients which are much essential for human beings. During floods, part of a river is carried by the simple main channel and rest is carried by flood plains. For such compound asymmetric channels, the flow structure becomes complicated due to momentum exchange between main channel and adjoining flood plains. Distribution of boundary shear in subsections provides us with the concept of momentum transfer between the interface of main channel and the flood plains. Experimentally, to get better data with accurate results are very complex because of the complexity of the problem. Hence, Conveyance Estimation System (CES) software has been used to tackle the complex processes to determine the shear stresses at different sections of an open channel having asymmetric flood plains on both sides of the main channel and the results are compared with the symmetric flood plains for various geometrical shapes and flow conditions. Error analysis is also performed to know the degree of accuracy of the model implemented.

Keywords: depth average velocity, non prismatic compound channel, relative flow depth , velocity distribution

Procedia PDF Downloads 108

Authors: Shogo Nagano, Yusuke Tanigaki, Hirokazu Fukuda

Abstract:

Vegetables harvested early in the morning or late in the afternoon are valued in plant production, and so the time of harvest is important. The biological functions known as circadian clocks have a significant effect on this harvest timing. The purpose of this study was to non-destructively estimate the circadian clock and so construct a method for determining a suitable harvest time. We took eight samples of green busil (Perilla frutescens var. crispa) every 4 hours, six times for 1 day and analyzed all samples at the same time. A hyperspectral camera was used to collect spectrum intensities at 141 different wavelengths (350–1050 nm). Calculation of correlations between spectrum intensity of each wavelength and harvest time suggested the suitability of the hyperspectral camera for non-destructive estimation. However, even the highest correlated wavelength had a weak correlation, so we used machine learning to raise the accuracy of estimation and constructed a machine learning model to estimate the internal time of the circadian clock. Artificial neural networks (ANN) were used for machine learning because this is an effective analysis method for large amounts of data. Using the estimation model resulted in an error between estimated and real times of 3 min. The estimations were made in less than 2 hours. Thus, we successfully demonstrated this method of non-destructively estimating internal time.

Keywords: artificial neural network (ANN), circadian clock, green busil, hyperspectral camera, non-destructive evaluation

Procedia PDF Downloads 285

Authors: Bingchun Liu, Pei-Chann Chang, Natasha Huang, Dun Li

Abstract:

Machine Learning and Data Mining are the two important tools for extracting useful information and knowledge from large datasets. In machine learning, classification is a wildly used technique to predict qualitative variables and is generally preferred over regression from an operational point of view. Due to the enormous increase in air pollution in various countries especially China, Air Quality Classification has become one of the most important topics in air quality research and modelling. This study aims at introducing a hybrid classification model based on information theory and Support Vector Machine (SVM) using the air quality data of four cities in China namely Beijing, Guangzhou, Shanghai and Tianjin from Jan 1, 2014 to April 30, 2016. China's Ministry of Environmental Protection has classified the daily air quality into 6 levels namely Serious Pollution, Severe Pollution, Moderate Pollution, Light Pollution, Good and Excellent based on their respective Air Quality Index (AQI) values. Using the information theory, information gain (IG) is calculated and feature selection is done for both categorical features and continuous numeric features. Then SVM Machine Learning algorithm is implemented on the selected features with cross-validation. The final evaluation reveals that the IG and SVM hybrid model performs better than SVM (alone), Artificial Neural Network (ANN) and K-Nearest Neighbours (KNN) models in terms of accuracy as well as complexity.

Keywords: machine learning, air quality classification, air quality index, information gain, support vector machine, cross-validation

Procedia PDF Downloads 222

Authors: Ejaz Ahmed, Huang Yong

Abstract:

The operation of aero engines has a critical importance in the vicinity of lean blowout (LBO) limits. Lefebvre’s model of LBO based on empirical correlation has been extended to flame volume concept by the authors. The flame volume takes into account the effects of geometric configuration, the complex spatial interaction of mixing, turbulence, heat transfer and combustion processes inside the gas turbine combustion chamber. For these reasons, flame volume based LBO predictions are more accurate. Although LBO prediction accuracy has improved, it poses a challenge associated with Vf estimation in real gas turbine combustors. This work extends the approach of flame volume prediction previously based on fuel iterative approximation with cold flow simulations to reactive flow simulations. Flame volume for 11 combustor configurations has been simulated and validated against experimental data. To make prediction methodology robust as required in the preliminary design stage, reactive flow simulations were carried out with the combination of probability density function (PDF) and discrete phase model (DPM) in FLUENT 15.0. The criterion for flame identification was defined. Two important parameters i.e. critical injection diameter (Dp,crit) and critical temperature (Tcrit) were identified, and their influence on reactive flow simulation was studied for Vf estimation. Obtained results exhibit ±15% error in Vf estimation with experimental data.

Keywords: CFD, combustion, gas turbine combustor, lean blowout

Procedia PDF Downloads 254

Authors: Stefan W. Kienzle

Abstract:

The increasing number of occurrences of extreme weather and climate events have significant impacts on society and are the cause of continued and increasing loss of human and animal lives, loss or damage to property (houses, cars), and associated stresses to the public in coping with a changing climate. A climate index breaks down daily climate time series into meaningful derivatives, such as the annual number of frost days. Climate indices allow for the spatially consistent analysis of a wide range of climate-dependent variables, which enables the quantification and mapping of historical and future climate change across regions. As trends of phenomena such as the length of the growing season change differently in different hydro-climatological regions, mapping needs to be carried out at a high spatial resolution, such as the 10km by 10km Canadian Climate Grid, which has interpolated daily values from 1950 to 2017 for minimum and maximum temperature and precipitation. Climate indices form the basis for the analysis and comparison of means, extremes, trends, the quantification of changes, and their respective confidence levels. A total of 39 temperature indices and 16 precipitation indices were computed for the period 1951 to 2017 for the Province of Alberta. Temperature indices include the annual number of days with temperatures above or below certain threshold temperatures (0, +-10, +-20, +25, +30ºC), frost days, and timing of frost days, freeze-thaw days, growing or degree days, and energy demands for air conditioning and heating. Precipitation indices include daily and accumulated 3- and 5-day extremes, days with precipitation, period of days without precipitation, and snow and potential evapotranspiration. The rank-based nonparametric Mann-Kendall statistical test was used to determine the existence and significant levels of all associated trends. The slope of the trends was determined using the non-parametric Sen’s slope test. The Google mapping interface was developed to create the website albertaclimaterecords.com, from which beach of the 55 climate indices can be queried for any of the 6833 grid cells that make up Alberta. In addition to the climate indices, climate normals were calculated and mapped for four historical 30-year periods and one future period (1951-1980, 1961-1990, 1971-2000, 1981-2017, 2041-2070). While winters have warmed since the 1950s by between 4 - 5°C in the South and 6 - 7°C in the North, summers are showing the weakest warming during the same period, ranging from about 0.5 - 1.5°C. New agricultural opportunities exist in central regions where the number of heat units and growing degree days are increasing, and the number of frost days is decreasing. While the number of days below -20ºC has about halved across Alberta, the growing season has expanded by between two and five weeks since the 1950s. Interestingly, both the number of days with heat waves and cold spells have doubled to four-folded during the same period. This research demonstrates the enormous potential of using climate indices at the best regional spatial resolution possible to enable society to understand historical and future climate changes of their region.

Keywords: climate change, climate indices, habitat risk, regional, mapping, extremes

Procedia PDF Downloads 80

Authors: Branislav Timotijevic, Yves Petremand, Markus Luetzelschwab, Dara Bayat, Laurent Aebi

Abstract:

Optical miniaturized sensors with remote readout are required devices for the monitoring in harsh electromagnetic environments. As an example, in turbo and hydro generators, excessively high vibrations of the end-windings can lead to dramatic damages, imposing very high, additional service costs. A significant change of the generator temperature can also be an indicator of the system failure. Continuous monitoring of vibrations, temperature, humidity, and gases is therefore mandatory. The high electromagnetic fields in the generators impose the use of non-conductive devices in order to prevent electromagnetic interferences and to electrically isolate the sensing element to the electronic readout. Metal-free sensors are good candidates for such systems since they are immune to very strong electromagnetic fields and given the fact that they are non-conductive. We have realized miniature optical accelerometer and temperature sensors for a remote sensing of the harsh environments using the common, inexpensive silicon Micro Electro-Mechanical System (MEMS) platform. Both devices show highly linear response. The accelerometer has a deviation within 1% from the linear fit when tested in a range 0 – 40 g. The temperature sensor can provide the measurement accuracy better than 1 °C in a range 20 – 150 °C. The design of other type of sensors for the environments with high electromagnetic interferences has also been discussed.

Keywords: optical MEMS, temperature sensor, accelerometer, remote sensing, harsh environment

Procedia PDF Downloads 352

Authors: Moschos Vogiatzis, K. Perakis

Abstract:

Spatial-explicit and up-to-date land use/land cover information is fundamental for spatial planning, land management, sustainable development, and sound decision-making. In the last decade, many satellite-derived land cover products at different spatial, spectral, and temporal resolutions have been developed, such as the European Copernicus Land Cover product. However, more efficient and detailed information for land use/land cover is required at the regional or local scale. A typical Mediterranean basin with a complex landscape comprised of various forest types, crops, artificial surfaces, and wetlands was selected to test and develop our approach. In this study, we investigate the improvement of Copernicus Land Cover product (CLC2018) using Landsat 8 and Sentinel-2 pixel-based classification based on all available existing geospatial data (Forest Maps, LPIS, Natura2000 habitats, cadastral parcels, etc.). We examined and compared the performance of the Random Forest classifier for land use/land cover mapping. In total, 10 land use/land cover categories were recognized in Landsat 8 and 11 in Sentinel-2A. A comparison of the overall classification accuracies for 2018 shows that Landsat 8 classification accuracy was slightly higher than Sentinel-2A (82,99% vs. 80,30%). We concluded that the main land use/land cover types of CLC2018, even within a heterogeneous area, can be successfully mapped and updated according to CLC nomenclature. Future research should be oriented toward integrating spatiotemporal information from seasonal bands and spectral indexes in the classification process.

Keywords: classification, land use/land cover, mapping, random forest

Procedia PDF Downloads 111

Authors: Jie Huo, Jonathan Wu

Abstract:

Segmentation of anatomical structures in medical images is essential for scientific inquiry into the complex relationships between biological structure and clinical diagnosis, treatment and assessment. As a method of incorporating the prior knowledge and the anatomical structure similarity between a target image and atlases, multi-atlas segmentation has been successfully applied in segmenting a variety of medical images, including the brain, cardiac, and abdominal images. The basic idea of multi-atlas segmentation is to transfer the labels in atlases to the coordinate of the target image by matching the target patch to the atlas patch in the neighborhood. However, this technique is limited by the pairwise registration between target image and atlases. In this paper, a novel multi-atlas segmentation approach is proposed by introducing a dynamic energy model. First, the target is mapped to each atlas image by minimizing the dynamic energy function, then the segmentation of target image is generated by weighted fusion based on the energy. The method is tested on MICCAI 2012 Multi-Atlas Labeling Challenge dataset which includes 20 target images and 15 atlases images. The paper also analyzes the influence of different parameters of the dynamic energy model on the segmentation accuracy and measures the dice coefficient by using different feature terms with the energy model. The highest mean dice coefficient obtained with the proposed method is 0.861, which is competitive compared with the recently published method.

Keywords: brain MRI segmentation, dynamic energy model, multi-atlas segmentation, energy minimization

Procedia PDF Downloads 322

Authors: Josif Figueroa, Kush Bubbar, Greg Young-Morris

Abstract:

The reliability evaluation of composite generation and bulk transmission systems is crucial for ensuring a reliable supply of electrical energy to significant system load points. However, evaluating adequacy indices using probabilistic methods like sequential Monte Carlo Simulation can be computationally expensive. Despite this, it is necessary when time-varying and interdependent resources, such as renewables and energy storage systems, are involved. Recent advances in solving power network optimization problems and parallel computing have improved runtime performance while maintaining solution accuracy. This work introduces CompositeSystems, an open-source Composite System Reliability Evaluation tool developed in Julia™, to address the current deficiencies of commercial and non-commercial tools. This work introduces its design, validation, and effectiveness, which includes analyzing two different formulations of the Optimal Power Flow problem. The simulations demonstrate excellent agreement with existing published studies while improving replicability and reproducibility. Overall, the proposed tool can provide valuable insights into the performance of transmission systems, making it an important addition to the existing toolbox for power system planning.

Keywords: open-source software, composite system reliability, optimization methods, Monte Carlo methods, optimal power flow

Procedia PDF Downloads 57

Authors: Jie Zhao, Meng Su

Abstract:

Image recognition, as one of the most critical technologies in computer vision, works to help machine-like robotics understand a scene, that is, if deployed appropriately, will trigger the revolution in remote sensing and industry automation. With the developments of AI technologies, there are many prevailing and sophisticated neural networks as technologies developed for image recognition. However, computer vision platforms as hardware, supporting neural networks for image recognition, as crucial as the neural network technologies, need to be more congruently addressed as the research subjects. In contrast, different computer vision platforms are deterministic to leverage the performance of different neural networks for recognition. In this paper, three different computer vision platforms – Jetson Nano(with 4GB), a standalone laptop(with RTX 3000s, using CUDA), and Google Colab (web-based, using GPU) are explored and four prominent neural network architectures (including AlexNet, VGG(16/19), GoogleNet, and ResNet(18/34/50)), are investigated. In the context of pairwise usage between different computer vision platforms and distinctive neural networks, with the merits of recognition accuracy and time efficiency, the performances are evaluated. In the case study using public imageNets, our findings provide a nuanced perspective on optimizing image recognition tasks across Edge-AI platforms, offering guidance on selecting appropriate neural network structures to maximize performance under hardware constraints.

Keywords: alexNet, VGG, googleNet, resNet, Jetson nano, CUDA, COCO-NET, cifar10, imageNet large scale visual recognition challenge (ILSVRC), google colab

Procedia PDF Downloads 70

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 145

Authors: Jie Dong, Binjie Hu, Andrzej W Pacek, Xiaogang Yang, Nicholas J. Miles

Abstract:

The effect of the shape of the vessel bottom and the length of baffles on the velocity distributions in a turbulent and in a transitional flow has been simulated. The turbulent flow was simulated using standard k-ε model and simulation was verified using LES whereas transitional flow was simulated using only LES. It has been found that both the shape of tank bottom and the baffles’ length has significant effect on the flow pattern and velocity distribution below the impeller. In the dished bottom tank with baffles reaching the edge of the dish, the large rotating volume of liquid was formed below the impeller. Liquid in this rotating region was not fully mixing. A dead zone was formed here. The size and the intensity of circulation within this zone calculated by k-ε model and LES were practically identical what reinforces the accuracy of the numerical simulations. Both types of simulations also show that employing full-length baffles can reduce the size of dead zone formed below the impeller. The LES was also used to simulate the velocity distribution below the impeller in transitional flow and it has been found that secondary circulation loops were formed near the tank bottom in all investigated geometries. However, in this case the length of baffles has smaller effect on the volume of rotating liquid than in the turbulent flow.

Keywords: baffles length, dished bottom, dead zone, flow field

Procedia PDF Downloads 286

Authors: Yan Zhang

Abstract:

Predictive maintenance is a technique to predict when an in-service machine will fail so that maintenance can be planned in advance. Analytics-driven predictive maintenance is gaining increasing attention in many industries such as manufacturing, utilities, aerospace, etc., along with the emerging demand of Internet of Things (IoT) applications and the maturity of technologies that support Big Data storage and processing. This study aims to build an end-to-end analytics solution that includes both real-time machine condition monitoring and machine learning based predictive analytics capabilities. The goal is to showcase a general predictive maintenance solution architecture, which suggests how the data generated from field machines can be collected, transmitted, stored, and analyzed. We use a publicly available aircraft engine run-to-failure dataset to illustrate the streaming analytics component and the batch failure prediction component. We outline the contributions of this study from four aspects. First, we compare the predictive maintenance problems from the view of the traditional reliability centered maintenance field, and from the view of the IoT applications. When evolving to the IoT era, predictive maintenance has shifted its focus from ensuring reliable machine operations to improve production/maintenance efficiency via any maintenance related tasks. It covers a variety of topics, including but not limited to: failure prediction, fault forecasting, failure detection and diagnosis, and recommendation of maintenance actions after failure. Second, we review the state-of-art technologies that enable a machine/device to transmit data all the way through the Cloud for storage and advanced analytics. These technologies vary drastically mainly based on the power source and functionality of the devices. For example, a consumer machine such as an elevator uses completely different data transmission protocols comparing to the sensor units in an environmental sensor network. The former may transfer data into the Cloud via WiFi directly. The latter usually uses radio communication inherent the network, and the data is stored in a staging data node before it can be transmitted into the Cloud when necessary. Third, we illustrate show to formulate a machine learning problem to predict machine fault/failures. By showing a step-by-step process of data labeling, feature engineering, model construction and evaluation, we share following experiences: (1) what are the specific data quality issues that have crucial impact on predictive maintenance use cases; (2) how to train and evaluate a model when training data contains inter-dependent records. Four, we review the tools available to build such a data pipeline that digests the data and produce insights. We show the tools we use including data injection, streaming data processing, machine learning model training, and the tool that coordinates/schedules different jobs. In addition, we show the visualization tool that creates rich data visualizations for both real-time insights and prediction results. To conclude, there are two key takeaways from this study. (1) It summarizes the landscape and challenges of predictive maintenance applications. (2) It takes an example in aerospace with publicly available data to illustrate each component in the proposed data pipeline and showcases how the solution can be deployed as a live demo.

Keywords: Internet of Things, machine learning, predictive maintenance, streaming data

Procedia PDF Downloads 374

Authors: Michael G. Pantelyat

Abstract:

Fundamental basics of pure and applied research in the area of magneto-thermo-mechanical numerical analysis and design of innovative electromagnetic devices (modern induction heaters, novel thermoelastic actuators, rotating electrical machines, induction cookers, electrophysical devices) are elaborated. Thus, mathematical models of magneto-thermo-mechanical processes in electromagnetic devices taking into account main interactions of interrelated phenomena are developed. In addition, graphical representation of coupled (multiphysics) phenomena under consideration is proposed. Besides, numerical techniques for nonlinear problems solution are developed. On this base, effective numerical algorithms for solution of actual problems of practical interest are proposed, validated and implemented in applied 2D and 3D computer codes developed. Many applied problems of practical interest regarding modern electrical engineering devices are numerically solved. Investigations of the influences of various interrelated physical phenomena (temperature dependences of material properties, thermal radiation, conditions of convective heat transfer, contact phenomena, etc.) on the accuracy of the electromagnetic, thermal and structural analyses are conducted. Important practical recommendations on the choice of rational structures, materials and operation modes of electromagnetic devices under consideration are proposed and implemented in industry.

Keywords: electromagnetic devices, multiphysics, numerical analysis, simulation and design

Procedia PDF Downloads 375

Authors: Saw Thiha, Jay Rajasekera

Abstract:

A boom in video conferencing generated millions of hours of video data daily to be analyzed. However, such enormous data pose certain scalability issues to be analyzed efficiently, let alone do it in real-time, as online conferences can involve hundreds of people and can last for hours. This paper proposes an efficient online proctoring solution that can analyze the online conferences real-time on edge devices such as Android, iOS, and desktops. Since the computation can be done upfront on the devices where online conferences take place, it can scale well without requiring intensive resources such as GPU servers and complex cloud infrastructure. According to the linear models, face orientation does indeed impact the perceived eye openness. Also, the proposed z score facial landmark standardization was proven to be functional in detecting face orientation and contributed to classifying eye blinks with single eyelid distance computation while achieving a better f1 score and accuracy than the Eye Aspect Ratio (EAR) threshold method. Last but not least, the authors implemented the solution natively in the MediaPipe framework and open-sourced it along with the reproducible experimental results on GitHub. The solution provides face orientation, eye blink, facial activity, and translation detections out of the box and is highly customizable and extensible.

Keywords: android, desktop, edge computing, blink, face orientation, facial activity and translation, MediaPipe, open source, real-time, video conference, web, iOS, Z score facial landmark standardization

Procedia PDF Downloads 83

Authors: Wannakarn Likitrattanaporn

Abstract:

The purposes of this investigation were to investigate the effects of applying linguistic principles and teaching techniques in teaching English through experimenting the Adapted English Lessons and to determine the teachers’ opinions as well as students’ opinions towards the Adapted Lessons. The subjects of the study were 5 Thai teachers, who teach English, and 85 Grade 10 mixed-ability students at Triamudom Suksa Pattanakarn Ratchada School, Bangkok, Thailand. The research instruments included the Adapted English Lessons, questionnaires asking teachers’ and students’ opinions towards the Adapted Lessons and the informal interview. The data from the research instruments was collected and analyzed concerning the teachers’ and students’ opinions towards adapting linguistic principles and teaching techniques. Linguistic principles of minimal pair and articulatory phonetics and teaching techniques of mimicry-memorization; vocabulary substitution drills, language pattern drills, reading comprehension exercise, practicing listening, speaking and writing skill and communicative activities; informal talk and free writing are applied. The data was statistically compiled according to an arithmetic percentage. The results showed that the teachers and students have very highly positive opinions towards adapting linguistic principles for teaching and learning phonological accuracy. Teaching techniques provided in the Adapted English Lessons can be used efficiently in the classroom. The teachers and students have positive opinions towards them too.

Keywords: applying linguistic principles and teaching techniques, teachers’ and students’ opinions, teaching English, the adapted English lessons

Procedia PDF Downloads 466

Authors: C. Manjula, Lilly Florence

Abstract:

Software technology is developing rapidly which leads to the growth of various industries. Now-a-days, software-based applications have been adopted widely for business purposes. For any software industry, development of reliable software is becoming a challenging task because a faulty software module may be harmful for the growth of industry and business. Hence there is a need to develop techniques which can be used for early prediction of software defects. Due to complexities in manual prediction, automated software defect prediction techniques have been introduced. These techniques are based on the pattern learning from the previous software versions and finding the defects in the current version. These techniques have attracted researchers due to their significant impact on industrial growth by identifying the bugs in software. Based on this, several researches have been carried out but achieving desirable defect prediction performance is still a challenging task. To address this issue, here we present a machine learning based hybrid technique for software defect prediction. First of all, Genetic Algorithm (GA) is presented where an improved fitness function is used for better optimization of features in data sets. Later, these features are processed through Decision Tree (DT) classification model. Finally, an experimental study is presented where results from the proposed GA-DT based hybrid approach is compared with those from the DT classification technique. The results show that the proposed hybrid approach achieves better classification accuracy.

Keywords: decision tree, genetic algorithm, machine learning, software defect prediction

Procedia PDF Downloads 320

Authors: Rongbo Shen, Jianhua Yao, Kezhou Yan, Kuan Tian, Cheng Jiang, Ke Zhou

Abstract:

This study proposes a generative adversarial networks (GAN) framework to perform synthetic sampling in feature space, i.e., feature augmentation, to address the class imbalance problem in medical image analysis. A feature extraction network is first trained to convert images into feature space. Then the GAN framework incorporates adversarial learning to train a feature generator for the minority class through playing a minimax game with a discriminator. The feature generator then generates features for minority class from arbitrary latent distributions to balance the data between the majority class and the minority class. Additionally, a data cleaning technique, i.e., Tomek link, is employed to clean up undesirable conflicting features introduced from the feature augmentation and thus establish well-defined class clusters for the training. The experiment section evaluates the proposed method on two medical image analysis tasks, i.e., mass classification on mammogram and cancer metastasis classification on histopathological images. Experimental results suggest that the proposed method obtains superior or comparable performance over the state-of-the-art counterparts. Compared to all counterparts, our proposed method improves more than 1.5 percentage of accuracy.

Keywords: class imbalance, synthetic sampling, feature augmentation, generative adversarial networks, data cleaning

Procedia PDF Downloads 114

Authors: Ajai Singh

Abstract:

Simulation of hydrological processes at the watershed outlet through modelling approach is essential for proper planning and implementation of appropriate soil conservation measures in Damodar Barakar catchment, Hazaribagh, India where soil erosion is a dominant problem. This study quantifies the parametric uncertainty involved in simulation of stream flow using Soil and Water Assessment Tool (SWAT), a watershed scale model and Radial Basis Neural Network (RBNN), an artificial neural network model. Both the models were calibrated and validated based on measured stream flow and quantification of the uncertainty in SWAT model output was assessed using ‘‘Sequential Uncertainty Fitting Algorithm’’ (SUFI-2). Though both the model predicted satisfactorily, but RBNN model performed better than SWAT with R2 and NSE values of 0.92 and 0.92 during training, and 0.71 and 0.70 during validation period, respectively. Comparison of the results of the two models also indicates a wider prediction interval for the results of the SWAT model. The values of P-factor related to each model shows that the percentage of observed stream flow values bracketed by the 95PPU in the RBNN model as 91% is higher than the P-factor in SWAT as 87%. In other words the RBNN model estimates the stream flow values more accurately and with less uncertainty. It could be stated that RBNN model based on simple input could be used for estimation of monthly stream flow, missing data, and testing the accuracy and performance of other models.

Keywords: SWAT, RBNN, SUFI 2, bootstrap technique, stream flow, simulation

Procedia PDF Downloads 349

Authors: Rohit Semwal, Shivam Arora, Saurav, Sangita Roy

Abstract:

This work proposes a supervised framework with deep convolutional neural networks (CNNs) for vision-based sign language recognition of static gestures. Our approach addresses the acquisition and segmentation of correct inputs for the CNN-based classifier. Microsoft Kinect™ sensor, despite complex environmental conditions, can track hands efficiently. Skin Colour based segmentation is applied on cropped images of hands in different poses, used to depict different sign language gestures. The segmented hand images are used as an input for our classifier. The CNN classifier proposed in the paper is able to classify the input images with a high degree of accuracy. The system was trained and tested on 39 static sign language gestures, including 26 letters of the alphabet and 13 commonly used words. This paper includes a problem definition for building the proposed system, which acts as a sign language translator between deaf/mute and the rest of the society. It is then followed by a focus on reviewing existing knowledge in the area and work done by other researchers. It also describes the working principles behind different components of CNNs in brief. The architecture and system design specifications of the proposed system are discussed in the subsequent sections of the paper to give the reader a clear picture of the system in terms of the capability required. The design then gives the top-level details of how the proposed system meets the requirements.

Keywords: sign language, CNN, HCI, segmentation

Procedia PDF Downloads 137

Authors: Ursula Das, Manavjit Singh Dhindsa, Kshirasagar Naik, Marzia Zaman, Richard Purcell, Srinivas Sampalli, Abdul Mutakabbir, Chung-Horng Lung, Thambirajah Ravichandran

Abstract:

Wildland fires, also known as forest fires or wildfires, are exhibiting an alarming surge in frequency in recent times, further adding to its perennial global concern. Forest fires often lead to devastating consequences ranging from loss of healthy forest foliage and wildlife to substantial economic losses and the tragic loss of human lives. Despite the existence of substantial literature on the detection of active forest fires, numerous potential research avenues in forest fire management, such as preventative measures and ancillary effects of forest fires, remain largely underexplored. This paper undertakes a systematic review of these underexplored areas in forest fire research, meticulously categorizing them into distinct phases, namely pre-fire, during-fire, and post-fire stages. The pre-fire phase encompasses the assessment of fire risk, analysis of fuel properties, and other activities aimed at preventing or reducing the risk of forest fires. The during-fire phase includes activities aimed at reducing the impact of active forest fires, such as the detection and localization of active fires, optimization of wildfire suppression methods, and prediction of the behavior of active fires. The post-fire phase involves analyzing the impact of forest fires on various aspects, such as the extent of damage in forest areas, post-fire regeneration of forests, impact on wildlife, economic losses, and health impacts from byproducts produced during burning. A comprehensive understanding of the three stages is imperative for effective forest fire management and mitigation of the impact of forest fires on both ecological systems and human well-being. Artificial intelligence and machine learning (AI/ML) methods have garnered much attention in the cyber-physical systems domain in recent times leading to their adoption in decision-making in diverse applications including disaster management. This paper explores the current state of AI/ML applications for managing the activities in the aforementioned phases of forest fire. While conventional machine learning and deep learning methods have been extensively explored for the prevention, detection, and management of forest fires, a systematic classification of these methods into distinct AI research domains is conspicuously absent. This paper gives a comprehensive overview of the state of forest fire research across more recent and prominent AI/ML disciplines, including big data, classical machine learning, computer vision, explainable AI, generative AI, natural language processing, optimization algorithms, and time series forecasting. By providing a detailed overview of the potential areas of research and identifying the diverse ways AI/ML can be employed in forest fire research, this paper aims to serve as a roadmap for future investigations in this domain.

Keywords: artificial intelligence, computer vision, deep learning, during-fire activities, forest fire management, machine learning, pre-fire activities, post-fire activities

Procedia PDF Downloads 56