Covid-19 and its impact on school closures: a predictive analysis using machine learning algorithms

Published: Aug. 5, 2021

This research presents an extensive point of reference for investigating the operation several machine learning (ML) algorithms in postulating multiclass classification problem regarding forthcoming effects Covid-19 on school closures. With prompt closure schools across world response to this pandemic, school-going children and teenagers are ruptured both mentally physically. Hence, ML has come be a reliable component forecast scenario effectively. A dataset from UNESCO is trained tested by ten supervised algorithms. comprehensive analysis among predictive models was executed which bought satisfactory results with regard accuracy, precision, sensitivity, F1 score, ROC-AUC hyper parameter optimization. In regard, grid search cross validation (GridSearchCV) utilized order obtain optimal parameters. However, performance Artificial Neural Network (ANN) also investigated compared where ANN displayed maximum accuracy 80.37%. After rigorous comparative analysis, Decision Tree (DT) portrayed highest 90.75%. it evident that algorithm holds strong promise forecasting upcoming closures due can contribute significantly decision making welfare education system.

Language: Английский

---

Accident Detection and Road Condition Monitoring Using Blackbox Module

2022 Thirteenth International Conference on Ubiquitous and Future Networks (ICUFN), Journal Year: 2022, Volume and Issue: unknown, P. 247 - 252

Published: July 5, 2022

Road accidents have always been a major cause of mass fatalities all over the world specially in Bangladesh, where conditions roads and whole traffic infrastructure general are less than ideal, collisions tend to occur more often. The influx road past few years has led researchers analyze accident. In this regard, smart system integrating recorded information like velocity, acceleration, rotation, position vehicle can play significant role detecting occurrence accident send prompt alerts first responders. This study aims introduce an approach with 'Blackbox' module that serves two purposes simultaneously reduce number accidents. mentioned equipped different sensors record important parameters as well overall condition while being capable properly collision. 'BlackBox' map data certain road's which be matched available particular location saved on database. Based data, alert provided concerned for prone locations. way, taking place each year reduced up 80%.

Language: Английский

---

Perceived Stress Analysis of Undergraduate Students during COVID-19: A Machine Learning Approach

2022 IEEE 21st Mediterranean Electrotechnical Conference (MELECON), Journal Year: 2022, Volume and Issue: unknown

Published: June 14, 2022

Stress is a state of mind when an individual experiences emotional or physical tensions originating from any event that results in frustration, anger, nervousness. Unfortunately, since the inception COVID-19 pandemic, it has been massively witnessed among university students due to persistent usage e-learning gears for last two years. Due severity observed stress, accurate and early prediction detection should play pivotal role treating student. In this work, questionnaire-based dataset on Jordanian analyzed using 5-point Likert Scale. One most widely used psychological instrument Perceived Scale (PSS) identify stress-related symptoms students. Based dataset, several machine learning (ML) algorithms were applied regression classification analysis by which mental stress predicted classified. After simulation Python, ML regressors evaluated through performance metrics such as R ² Score, Root Mean Squared Error (RMSE), Absolute (MAE), Percentage (MAPE), classifiers assessed accuracy, precision, recall, F1-Score. It Linear Regression (LR) performed best all models whereas Logistic Classifier (LRC) portrayed highest accuracy 97.8% classifiers. Therefore, ML-based can significantly contribute analyzing students' during automated manner.

Language: Английский

---

Machine Learning Assisted Decision Support System for Prediction of Prostrate Cancer

2022 19th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), Journal Year: 2023, Volume and Issue: unknown, P. 1 - 5

Published: May 9, 2023

Over the past several years, there has been a global rise in prevalence of prostate cancer. It was discovered that cancer is most often diagnosed category amongst men and it can be stated as main cause cancer-related mortality worldwide among males. Diagnosing illnesses one greatest obstacles medicine. This study crucial due to lack precise standards for evaluation symptoms low predictive accuracy current diagnostic approaches. believed machine learning approaches may used solve situations when are no defined rules where event-influencing aspects predicted. Computer-aided systems produce variety solutions with this knowledge. In study, performance various supervised algorithms (SVC, LR, AdaBoost (Ada B), XG Boost (XGB), KNC, LGBM, GB, DT, RF) compared discussed. we acquired data from Kaggle consisting 100 cases 10 characteristics. our model, initially determined maximum XGB, RF 93.33 percent. Eventually, GridsearchCV tune hyperparameters order improve classifiers. time, highest 96.67% not just those three, but also GB whole. The noteworthy finding improvement consistency predictions. Therefore, if computer educated methods using patient data, therapeutically beneficial predicting high degree accuracy. method, an unnecessary biopsy avoided.

Language: Английский

---

Classification of adaptive back propagation neural network along with fuzzy logic in chronic kidney disease

e-Prime - Advances in Electrical Engineering Electronics and Energy, Journal Year: 2024, Volume and Issue: 7, P. 100463 - 100463

Published: Feb. 9, 2024

A steady deterioration in kidney function over months or years is known as chronic disease (CKD). Through a range of techniques, such pharmacological intervention moderate cases and hemodialysis renal transport severe cases, early identification CKD crucial has substantial influence on reducing the patient's health development. The outcomes show kidneys' present state. It suggested to develop system for detecting using machine learning. Finding best feature sets typically involves metaheuristic algorithms since selection an NP-hard issue with amorphous polynomials. Semi-crystalline tabu search (TS) frequently used both local global searches. In this study, we employ brand-new hybrid TS stochastic diffusion (SDX)-based selection. adaptive backpropagation neural network (ABPNN-ANFIS) then classified fuzzy logic. Fuzzy logic may be combine ABPNN findings. Consequently, these techniques can aid experts determining stage disease. Adaptive Neuron Clearing Inference System was utilised inverse networks MATLAB programme. demonstrate that ABPNN-ANFIS 98% accurate terms efficiency.

Language: Английский

---

Detection of Autism Spectrum Disorder by Discriminant Analysis Algorithm

Lecture notes on data engineering and communications technologies, Journal Year: 2021, Volume and Issue: unknown, P. 473 - 482

Published: Dec. 3, 2021

Language: Английский

---

Detection of Epileptic Seizure from EEG Signal Data by Employing Machine Learning Algorithms with Hyperparameter Optimization

Published: Dec. 8, 2021

Epileptic seizure refers to a brief occurrence of signs in the brain caused by abnormally high or synchronized neuronal activity. With utilization EEG signal, epileptic can be identified. However, incorporating machine learning classifiers with this data significantly contribute detecting an automated manner. In paper, nine algorithms have been studied and models constructed utilizing UCI Seizure dataset. The performances ML are noted detailed comparative analysis has exhibited for both hyperparameter tuning without tuning. Random search cross validation used hyperparameters. Satisfactory results witnessed terms different performance metrics like accuracy, precision, recall, specificity, FI-Score, ROC. After simulation, Support Vector Machine (SVM) performed best accuracy over 97.86%. Forest (RF) Multi-Layer Perceptron (MLP) also depicted promising accuracies 97.50% 97.26% respectively. Therefore, proper implementation based diagnosis system, patients having seizures identified treated at early stage.

Language: Английский

---

An Investigative Approach to Employ Support Vector Classifier as a Potential Detector of Brain Cancer from MRI Dataset

2021 International Conference on Electronics, Communications and Information Technology (ICECIT), Journal Year: 2021, Volume and Issue: unknown, P. 1 - 4

Published: Sept. 14, 2021

This paper proposes an investigative analysis to study the applicability of support vector classifier (SVC) algorithm detect brain cancer efficiently. Brain cancer, mostly triggered by tumor cells in can be too lethal if malignancy is not identified at early stage. Timely and tailored treatment plan will lead optimistic result which lessen magnitude disease. But it really challenging figure out malignancies manually from a large MRI dataset. In this context, dataset kaggle has been deployed conduct multiclass classification SVC where information extracted pictures. However, Linear NuSVC are also investigated apart traditional method. order increase performance models, grid search cross validation applied tune hyperparameters. All confusion matrices for both tuning without hyperparameters presented comprehensive manner thus, parameters tabulated evaluated extensively. Among three approaches, depicts maximum accuracy 95.71% along with

Language: Английский

---

Real-Time Clinical Gait Analysis and Foot Anomalies Detection Using Pressure Sensors and Convolutional Neural Network

2022 7th International Conference on Business and Industrial Research (ICBIR), Journal Year: 2022, Volume and Issue: unknown, P. 717 - 722

Published: May 19, 2022

This research presents a novel insight on gait disorder detection using transfer learning algorithms sensor-acquired data based the implementation of popular Convolutional Neural Network (CNN) models. The paper proposes use pressure sensors to extract heatmap images during gait, which are then trained and tested in various classification for abnormality diagnosis detection. Gait is biological scientific study body movement locomotion that emphatically serves as reliable parameter inspecting human body's neuromuscular skeletal systems. To build convenient precise system possible application, synthetic was generated multiple preexisting CNN models, were evaluated conventional performance metrics. proposed notion yielded experimental findings showed higher accuracies all schemes tested, with Vgg16 model achieving notable accuracy 97.15%. As result, analysis demonstrated not only significant terms accuracy, but also reduced complexity computing time, making approach efficient yet effective.

Language: Английский

---

An Efficient Analysis of EEG Signals to Perform Emotion Analysis

Published: May 9, 2023

The analysis of human emotional features is a significant hurdle to surmount on the path understanding mind. Human emotions are convoluted thus making its even more daunting. In this paper, meticulous and thorough EEG Brainwave Dataset: Feeling Emotions performed in order classify three basic sentiments experienced by people. A Machine Learning (ML) based framework proposed execute multi-class classification process identify positive, neutral negative experiences dataset analysed two distinct ways. first method employs chi-square algorithm select 500 best from each sample which then employed classifying multiple utilizing several machine learning models. second utilizes sparsePCA for feature extraction before conducting with help It supplemented binary individual available entire analyze efficacy these ML algorithms-Support Vector Machines (SVM), Random Forest (RF), Light Gradient Boosting (LGBM) Multi-Layer Perceptron (MLP) investigative study. Maximum accuracy 99.25% precision obtained application LGBM model after optimization hyper-parameters used extraction.

Language: Английский

---