Pub Date : 2021-12-01DOI: 10.1109/ComPE53109.2021.9752120
S. Bhowmik, A. Mitra, P. Deb
In this proposed work, a mathematical model of transformer has been implemented using Simulink environment to investigate the effect of load switching on the induced e.m.f. due to the presence of leakage inductances of the windings. Estimation of the parameters on a laboratory single-phase transformer has been carried out through Open Circuit (O.C.) and Short Circuit (S.C.) tests and the transient equation for no-load current has been established. This experimental work defines the switching effect of resistive, resistive-inductive (RL) along with a power factor improvement capacitor introduced in parallel with RL load. It has been found in presence of capacitor, the inductive voltage spikes is minimized because of reactive power injection to the system.
{"title":"Effect of Load Switching on Induced e.m.f. of a Transformer","authors":"S. Bhowmik, A. Mitra, P. Deb","doi":"10.1109/ComPE53109.2021.9752120","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9752120","url":null,"abstract":"In this proposed work, a mathematical model of transformer has been implemented using Simulink environment to investigate the effect of load switching on the induced e.m.f. due to the presence of leakage inductances of the windings. Estimation of the parameters on a laboratory single-phase transformer has been carried out through Open Circuit (O.C.) and Short Circuit (S.C.) tests and the transient equation for no-load current has been established. This experimental work defines the switching effect of resistive, resistive-inductive (RL) along with a power factor improvement capacitor introduced in parallel with RL load. It has been found in presence of capacitor, the inductive voltage spikes is minimized because of reactive power injection to the system.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123406733","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-12-01DOI: 10.1109/ComPE53109.2021.9751830
Abhishek Singh, A. Payal
Recent advancements in various ICT technologies and integration of machine-learning and AI have further extended the Internet of Things (IoT) perspective from ’passive system of interconnected things’ to "an active multi-agent-based system of interconnected everything". However, true realization of future IoT solutions requires addressing critical challenges including limited network coverage and resources. Unmanned Aerial Vehicles (UAVs) have recently gained significant attention due to their acute mobility, equitable operational costs, flexible deployment, and autonomous capabilities. Efforts are being made towards integrating drones in IoT as a solution to the critical challenges. In this paper, we highlighted the key expectations based on the new perspective of IoT and summarized the challenges in IoT due to its inherent nature towards contentment of those expectations. Finally, we investigated the roles of UAVs at various functional layers of IoT-workflow architecture towards addressing the critical issues and enabling key expectations in future-IoT solutions.
{"title":"Roles of UAVs in IoT work-flow architecture","authors":"Abhishek Singh, A. Payal","doi":"10.1109/ComPE53109.2021.9751830","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9751830","url":null,"abstract":"Recent advancements in various ICT technologies and integration of machine-learning and AI have further extended the Internet of Things (IoT) perspective from ’passive system of interconnected things’ to \"an active multi-agent-based system of interconnected everything\". However, true realization of future IoT solutions requires addressing critical challenges including limited network coverage and resources. Unmanned Aerial Vehicles (UAVs) have recently gained significant attention due to their acute mobility, equitable operational costs, flexible deployment, and autonomous capabilities. Efforts are being made towards integrating drones in IoT as a solution to the critical challenges. In this paper, we highlighted the key expectations based on the new perspective of IoT and summarized the challenges in IoT due to its inherent nature towards contentment of those expectations. Finally, we investigated the roles of UAVs at various functional layers of IoT-workflow architecture towards addressing the critical issues and enabling key expectations in future-IoT solutions.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123925361","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-12-01DOI: 10.1109/ComPE53109.2021.9751877
M. Kumar, S. Das
Two specific conditions, such as maximum torque per inverter ampere (MTPIA) and unity primary power factor (UPPF) are considered in the present work for a comparative performance analysis while speed control of brushless doubly-fed reluctance generator (BDFRG) using primary field-oriented control (PFOC). The study is based on the active power, reactive power, and power factor of both the stator windings of BDFRG in super-synchronous, synchronous, and sub-synchronous speed zones. The study also deals with the assessment of the minimum rating of the inverter required in both the conditions for successful operations. The relevant studies are done in MATLAB/Simulink. The prima facie objective of the present work is to affirm the candidature of BDFRG in wind power generation.
{"title":"Speed Control of Brushless Doubly-fed Reluctance Generator under MTPIA and UPPF Conditions for Wind Power Application","authors":"M. Kumar, S. Das","doi":"10.1109/ComPE53109.2021.9751877","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9751877","url":null,"abstract":"Two specific conditions, such as maximum torque per inverter ampere (MTPIA) and unity primary power factor (UPPF) are considered in the present work for a comparative performance analysis while speed control of brushless doubly-fed reluctance generator (BDFRG) using primary field-oriented control (PFOC). The study is based on the active power, reactive power, and power factor of both the stator windings of BDFRG in super-synchronous, synchronous, and sub-synchronous speed zones. The study also deals with the assessment of the minimum rating of the inverter required in both the conditions for successful operations. The relevant studies are done in MATLAB/Simulink. The prima facie objective of the present work is to affirm the candidature of BDFRG in wind power generation.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128373502","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-12-01DOI: 10.1109/ComPE53109.2021.9752099
S. Khanday, Hoor Fatima, N. Rakesh
During the Covid-19 pandemic world has witnessed the rise of cyber-attacks, especially during the Lockdown time course announced by the countries throughout the world, when almost every aspect of life changed the routine from offline to online. Protecting and securing information resources during pandemics has been a top priority for the modern computing world, with databases, banking, E-commerce and mailing services, etc. being the eye-catching credentials to the attackers. Apart from cryptography, machine learning and deep learning can offer an enormous amount of help in testing, training, and extracting negligible information from the data sets. Deep learning and machine learning have many methods and models in the account to detect and classify the different versions of cyber-attacks occasionally, from the datasets. Some of the most common deep learning methods inspired by the neural networks are Recurrent Neural Networks, Convolutional Neural Networks, Deep Belief Networks, Deep Boltzman Networks, Autoencoders, and Stacked Auto-encoders. Also counting machine learning algorithms into the account, there is a vast variety of algorithms that are meant to perform classification and regression. The survey will provide some of the most important deep learning and machine learning architectures used for Cyber-security and can offer protective services against cyber-attacks. The paper is a survey about various categories of cyber-attacks with a timeline of different attacks that took place in India and some of the other countries in the world. The final section of the report is about what deep learning methods can offer for developing and improving the security policies and examining vulnerabilities of an information system.
{"title":"Deep learning offering resilience from trending cyber-attacks, a review","authors":"S. Khanday, Hoor Fatima, N. Rakesh","doi":"10.1109/ComPE53109.2021.9752099","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9752099","url":null,"abstract":"During the Covid-19 pandemic world has witnessed the rise of cyber-attacks, especially during the Lockdown time course announced by the countries throughout the world, when almost every aspect of life changed the routine from offline to online. Protecting and securing information resources during pandemics has been a top priority for the modern computing world, with databases, banking, E-commerce and mailing services, etc. being the eye-catching credentials to the attackers. Apart from cryptography, machine learning and deep learning can offer an enormous amount of help in testing, training, and extracting negligible information from the data sets. Deep learning and machine learning have many methods and models in the account to detect and classify the different versions of cyber-attacks occasionally, from the datasets. Some of the most common deep learning methods inspired by the neural networks are Recurrent Neural Networks, Convolutional Neural Networks, Deep Belief Networks, Deep Boltzman Networks, Autoencoders, and Stacked Auto-encoders. Also counting machine learning algorithms into the account, there is a vast variety of algorithms that are meant to perform classification and regression. The survey will provide some of the most important deep learning and machine learning architectures used for Cyber-security and can offer protective services against cyber-attacks. The paper is a survey about various categories of cyber-attacks with a timeline of different attacks that took place in India and some of the other countries in the world. The final section of the report is about what deep learning methods can offer for developing and improving the security policies and examining vulnerabilities of an information system.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130402739","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-12-01DOI: 10.1109/ComPE53109.2021.9752239
N. Mohan, R. Murugan, Tripti Goel, Parthapratim Roy
Diabetic retinopathy (DR) is a chronic disease leading cause of blindness. One of the primary symptoms of DR is exudates (EX). The EX is a condition in which proteins, lipids, water leaked to retinal areas causes vision impairment. The two types of EX are hard EX and soft EX based on their appearance and leakage consistency. Early intervention of DR diminishes the likelihood of vision loss. Therefore, an automated technique is required. We present a novel U-Net model that detects both soft and hard EX in this paper. The proposed model is implemented in two stages. Preprocessing of fundus images is included in the first. The custom residual blocks-based designed network is the second phase. The model is tested on two benchmark databases available publicly IDRiD and e-Ophtha. The results achieved using the proposed approach are better than other approaches.
{"title":"Exudate Detection with Improved U-Net Using Fundus Images","authors":"N. Mohan, R. Murugan, Tripti Goel, Parthapratim Roy","doi":"10.1109/ComPE53109.2021.9752239","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9752239","url":null,"abstract":"Diabetic retinopathy (DR) is a chronic disease leading cause of blindness. One of the primary symptoms of DR is exudates (EX). The EX is a condition in which proteins, lipids, water leaked to retinal areas causes vision impairment. The two types of EX are hard EX and soft EX based on their appearance and leakage consistency. Early intervention of DR diminishes the likelihood of vision loss. Therefore, an automated technique is required. We present a novel U-Net model that detects both soft and hard EX in this paper. The proposed model is implemented in two stages. Preprocessing of fundus images is included in the first. The custom residual blocks-based designed network is the second phase. The model is tested on two benchmark databases available publicly IDRiD and e-Ophtha. The results achieved using the proposed approach are better than other approaches.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128754052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-12-01DOI: 10.1109/ComPE53109.2021.9752429
Pawan Jindal, V. Khemchandani, Sushil Chandra, Vishal Pandey
Creating environments and dangerous scenarios for physical training is very difficult and has a very high cost in terms of money and men’s power.Virtual Reality is a technology that simulates real-life experiences and allows people to don their own cyber avatars in a virtual world and interact with it like they would in the real world. The application of VR technology in the defence paradigm is to make trainees and officers better at using equipment, navigating a mode of transport, gaining experience of potential combat situations, medical training and more. One of the advantages of VR training in defence is that it offers the functionality to immerse users in a virtual yet safe world.Our immersive system provides an intuitive way for the users to interact with the VR or AR world by physically moving around the real world and aiming freely with tangible objects. This encourages physical interaction between the players as they compete or collaborate with other players. We present a new immersive multiplayer simulation game developed for defence training. We developed three game environments which are Combat situation, Bomb defusal, and Hostage rescue, and players can see their performance based on previously played games.
{"title":"A Multiplayer Shooting Game Based Simulation For Defence Training","authors":"Pawan Jindal, V. Khemchandani, Sushil Chandra, Vishal Pandey","doi":"10.1109/ComPE53109.2021.9752429","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9752429","url":null,"abstract":"Creating environments and dangerous scenarios for physical training is very difficult and has a very high cost in terms of money and men’s power.Virtual Reality is a technology that simulates real-life experiences and allows people to don their own cyber avatars in a virtual world and interact with it like they would in the real world. The application of VR technology in the defence paradigm is to make trainees and officers better at using equipment, navigating a mode of transport, gaining experience of potential combat situations, medical training and more. One of the advantages of VR training in defence is that it offers the functionality to immerse users in a virtual yet safe world.Our immersive system provides an intuitive way for the users to interact with the VR or AR world by physically moving around the real world and aiming freely with tangible objects. This encourages physical interaction between the players as they compete or collaborate with other players. We present a new immersive multiplayer simulation game developed for defence training. We developed three game environments which are Combat situation, Bomb defusal, and Hostage rescue, and players can see their performance based on previously played games.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126001107","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
With enormous and voluminous data being generated on a regular basis at an exponential speed, there is a demanding need for concise and relevant information to be available for the masses. Traditionally, lengthy textual contents are manually summarized by Linguists or Domain Experts, which are highly time consuming and unfairly biased. There is a dire need for Automatic Text Summarization approaches to be introduced in this broad spectrum. Extractive Summarization is one such approach where the salient information or excerpts are identified from a source and extracted to generate a concise summary. TextRank is an unsupervised extractive summarization technique incorporating graph-based ranking of extracted texts and finding the most relevant excerpts to generate a concise summary. In this paper, the prospects of a domain agnostic algorithm like TextRank for various domains of News Article Summarization are explored, exploring its efficiency in domain specific tasks and conveniently drawing various insights. NLP based pre-processing approaches and Static Word Embeddings were leveraged with semantic cosine similarity for the efficient ranking of textual data and performance evaluation on various domains of BBC News Articles Summarization datasets through ROUGE metrics. A commendable ROUGE score is achieved.
{"title":"Graph Based Extractive News Articles Summarization Approach leveraging Static Word Embeddings","authors":"Utpal Barman, Vishal Barman, Mustafizur Rahman, Nawaz Khan Choudhury","doi":"10.1109/ComPE53109.2021.9752056","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9752056","url":null,"abstract":"With enormous and voluminous data being generated on a regular basis at an exponential speed, there is a demanding need for concise and relevant information to be available for the masses. Traditionally, lengthy textual contents are manually summarized by Linguists or Domain Experts, which are highly time consuming and unfairly biased. There is a dire need for Automatic Text Summarization approaches to be introduced in this broad spectrum. Extractive Summarization is one such approach where the salient information or excerpts are identified from a source and extracted to generate a concise summary. TextRank is an unsupervised extractive summarization technique incorporating graph-based ranking of extracted texts and finding the most relevant excerpts to generate a concise summary. In this paper, the prospects of a domain agnostic algorithm like TextRank for various domains of News Article Summarization are explored, exploring its efficiency in domain specific tasks and conveniently drawing various insights. NLP based pre-processing approaches and Static Word Embeddings were leveraged with semantic cosine similarity for the efficient ranking of textual data and performance evaluation on various domains of BBC News Articles Summarization datasets through ROUGE metrics. A commendable ROUGE score is achieved.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127256057","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-12-01DOI: 10.1109/ComPE53109.2021.9752331
Pachunoori Anusha, Sanjib Patra, Ayanava Roy, D. Saha
This technical write-up emphasizes on combined regulation of frequency - voltage in a restructured renewable energy integrated power system under PoolCo and bilateral transactions. A two area coordinated modeling of area load frequency control (ALFC) and automatic voltage regulator (AVR) is carried out in presence of thermal, hydro, systems. Non linearities are incorporated to get a realistic insight with scheduled delay, rate constraint, and dead band. A powerful algorithm namely Firefly Algorithm based Industrial controllers serve the purpose of classical controller as Secondary control in the considered two area power system. Selection of best secondary controller among integral (I), proportional-integral (PI) and proportional-integral - derivative (PID) controller is carried out based on a fair comparison under Pool Co Transaction Scenario. PID controller serves better. Further investigations are carried out with excitation in AVR loop along with step load perturbation at ALFC in both control areas. PID outperforms I and PI in stabilizing the system responses such as frequency deviation, tie-power deviation and voltage deviations.
{"title":"Combined Frequency and Voltage Control of a Deregulated Hydro-Thermal Power System employing FA based Industrial Controller","authors":"Pachunoori Anusha, Sanjib Patra, Ayanava Roy, D. Saha","doi":"10.1109/ComPE53109.2021.9752331","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9752331","url":null,"abstract":"This technical write-up emphasizes on combined regulation of frequency - voltage in a restructured renewable energy integrated power system under PoolCo and bilateral transactions. A two area coordinated modeling of area load frequency control (ALFC) and automatic voltage regulator (AVR) is carried out in presence of thermal, hydro, systems. Non linearities are incorporated to get a realistic insight with scheduled delay, rate constraint, and dead band. A powerful algorithm namely Firefly Algorithm based Industrial controllers serve the purpose of classical controller as Secondary control in the considered two area power system. Selection of best secondary controller among integral (I), proportional-integral (PI) and proportional-integral - derivative (PID) controller is carried out based on a fair comparison under Pool Co Transaction Scenario. PID controller serves better. Further investigations are carried out with excitation in AVR loop along with step load perturbation at ALFC in both control areas. PID outperforms I and PI in stabilizing the system responses such as frequency deviation, tie-power deviation and voltage deviations.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126642870","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-12-01DOI: 10.1109/ComPE53109.2021.9751797
Atharva Bankar, Rishabh Shinde, S. Bhingarkar
Computer vision is a top-tier domain of the technological world that is responsible for automating the visual systems from healthcare to self-driving vehicles. With a reputation for surpassing human intelligence, it can be implemented in various trigger systems like wildfire smoke detection where the emission of smoke as a result of wildfire is fairly unpredictable.Low contrast and brightness have a detrimental effect on computer vision tasks. We present a novel approach to detect forest wildfire smoke, using image translation for converting nighttime images to day time which eliminates the confusion between smoke, cloud, and fog. This translation aids the YOLOv5 object detection algorithm to detect the smoke with the same aptness irrespective of time and lighting conditions. This paper demonstrates that the object detection model performs better on the images translated to day time with a better confidence score as compared to the corresponding nighttime images.
{"title":"Impact of Image Translation using Generative Adversarial Networks for Smoke Detection","authors":"Atharva Bankar, Rishabh Shinde, S. Bhingarkar","doi":"10.1109/ComPE53109.2021.9751797","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9751797","url":null,"abstract":"Computer vision is a top-tier domain of the technological world that is responsible for automating the visual systems from healthcare to self-driving vehicles. With a reputation for surpassing human intelligence, it can be implemented in various trigger systems like wildfire smoke detection where the emission of smoke as a result of wildfire is fairly unpredictable.Low contrast and brightness have a detrimental effect on computer vision tasks. We present a novel approach to detect forest wildfire smoke, using image translation for converting nighttime images to day time which eliminates the confusion between smoke, cloud, and fog. This translation aids the YOLOv5 object detection algorithm to detect the smoke with the same aptness irrespective of time and lighting conditions. This paper demonstrates that the object detection model performs better on the images translated to day time with a better confidence score as compared to the corresponding nighttime images.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122767258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-12-01DOI: 10.1109/ComPE53109.2021.9751945
Harsh Walia, J. S.
COVID-19, which is subsequently named as SARS-CoV-2, First Human case was found in the City of Wuhan, from China, in Dec 2019. After that, the World health organization (WHO) has declared Coronavirus as a Pandemic on 11th March 2020. In this study, our primary aim is to Detect the Severe Covid-19 patient in the Early Stages by looking at the information on admission laboratory values, demographics, comorbidities, admission medications, admission supplementary oxygen orders, discharge, and mortality. 4711 patient’s dataset with confirmed SARS-CoV-2 infections are included in the study. Each Patient has total of 85 Features in the Dataset. So, we have Filtered the Top Best 35 features out of 85 features from the Dataset using the seven different feature Selection algorithm and taken the most common features out from the different feature Selection algorithm. After selecting the top most essential features, we have applied around 17 different kinds of ML models like Linear Regression, Logistic regression, SVM, LinearSVC, MLP-Classifier, Decision Tree Classifier, Gradient Boosting Classifier, AdaBoost, Random Forest, XGBoost, LightGBM Classifier, Ridge Classifier, Bagging Classifier, ExtraTreeClassifier, KNN, Naive Bayes, Neural network with Keras, and finally, a Voting Classifier which is the ensemble of all the Top Models from the above-mentioned Models. Finally, all Models are Compared on the basis of Area under the receiver operating characteristic (AUC) & get the best AUC as 0.89.
COVID-19，后来被命名为SARS-CoV-2，于2019年12月在中国武汉市发现了首例人类病例。此后，世界卫生组织(世卫组织)于2020年3月11日宣布冠状病毒为大流行。在这项研究中，我们的主要目的是通过查看入院实验室值、人口统计学、合并症、入院药物、入院补充氧单、出院和死亡率等信息，在早期发现重症Covid-19患者。4711例确诊的SARS-CoV-2感染患者的数据集被纳入研究。每个患者在数据集中共有85个特征。因此，我们使用七种不同的特征选择算法从数据集中的85个特征中筛选出了最好的35个特征，并从不同的特征选择算法中提取了最常见的特征。在选择了最重要的特征后，我们应用了大约17种不同的ML模型，如线性回归，逻辑回归，SVM，线性svc, mlp分类器，决策树分类器，梯度增强分类器，AdaBoost，随机森林，XGBoost, LightGBM分类器，Ridge分类器，Bagging分类器，extratreecclassifier, KNN，朴素贝叶斯，神经网络与Keras，最后，一个投票分类器，它是上述模型中所有顶级模型的集合。最后，根据接收机工作特性下面积(Area under the receiver operating characteristic, AUC)对各模型进行比较，得到最佳AUC为0.89。
{"title":"Early Mortality Risk Prediction in Covid-19 Patients Using an Ensemble of Machine Learning Models","authors":"Harsh Walia, J. S.","doi":"10.1109/ComPE53109.2021.9751945","DOIUrl":"https://doi.org/10.1109/ComPE53109.2021.9751945","url":null,"abstract":"COVID-19, which is subsequently named as SARS-CoV-2, First Human case was found in the City of Wuhan, from China, in Dec 2019. After that, the World health organization (WHO) has declared Coronavirus as a Pandemic on 11th March 2020. In this study, our primary aim is to Detect the Severe Covid-19 patient in the Early Stages by looking at the information on admission laboratory values, demographics, comorbidities, admission medications, admission supplementary oxygen orders, discharge, and mortality. 4711 patient’s dataset with confirmed SARS-CoV-2 infections are included in the study. Each Patient has total of 85 Features in the Dataset. So, we have Filtered the Top Best 35 features out of 85 features from the Dataset using the seven different feature Selection algorithm and taken the most common features out from the different feature Selection algorithm. After selecting the top most essential features, we have applied around 17 different kinds of ML models like Linear Regression, Logistic regression, SVM, LinearSVC, MLP-Classifier, Decision Tree Classifier, Gradient Boosting Classifier, AdaBoost, Random Forest, XGBoost, LightGBM Classifier, Ridge Classifier, Bagging Classifier, ExtraTreeClassifier, KNN, Naive Bayes, Neural network with Keras, and finally, a Voting Classifier which is the ensemble of all the Top Models from the above-mentioned Models. Finally, all Models are Compared on the basis of Area under the receiver operating characteristic (AUC) & get the best AUC as 0.89.","PeriodicalId":211704,"journal":{"name":"2021 International Conference on Computational Performance Evaluation (ComPE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129590183","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：