Authors: Nisha Sawant, Dnyandev Ravindra Khadapkar
DOI Link: https://doi.org/10.22214/ijraset.2022.48311
Certificate: View Certificate
Most educational institutions worldwide have been closed since March 2020 in an effort to slow the spread of the Covid-19 epidemic. More than 90% of students around the world have been influenced by this. In this study, we\'ll make a prediction about whether or not the Covid-19 epidemic has benefited student performance. Our data will be divided into training and testing datasets, with 80% of the data utilised for training and 20% for testing. To calculate the accuracy of our predictions, we\'ll use six different algorithms, including the RandomForestClassifier Algorithm, the GaussianNB Algorithm, the K Neighbors Classifier Algorithm, the Logistic Regression Algorithm, the Linear Discriminant Analysis Algorithm, and the DecisionTree Classifier Algorithm.
A. RandomForestClassifier: Suitable for Binary, Continuous and categorical data type.
The Random Forest Algorithm consists of several decision trees on various subsets of a given dataset. Based on the concept of ensemble learning process, it creates decision trees based on data samples. It gets the prediction from each of them and selects the best solution by means of voting.
B. LogisticRegression: Suitable for Binary data type.
Logistical Regression is a statistical method used for building machine learning models. It is considered as one of the simplest Machine Learning Algorithm which can be applied on various classification problems such as Spam Detection, Diabetes, Prediction, Cancer detection etc. There are only two possible outcomes in a logistic Regression formula(Dichotomous).
C. LinearDiscriminantAnalysis: Suitable for Independent variables
The LDA is a supervised algorithm that aims to find the linear discriminant to represent the axes that maximize separation between different classes of data. This reduces the number of features which reduces the computing cost significantly (similar to principal component analysis). It is used in face recognition, prediction, customer identification, medical fields etc.
D. KNeighbours: Suitable for Individual Data.
The K nearest Neighbour is useful when you are performing a pattern recognition test. It classifies a data point based on it's neighbor's classification and stores all available cases. Although it is mostly recommended for Classification problems, it can also be used for regression. The algorithm is Non-Parametric, which means it does not make any assumption on underlying data.
A decision tree is a graphical representation for getting all the possible solutions to a problem/decision based on given conditions. They can be used in both regression and classification tasks. A decision tree comprises of two nodes, a decision node and Leaf Node. Decision nodes are used to make any decision and have multiple branches, whereas Leaf Nodes are the output of these decisions.
F. GaussianNB: Suitable for Continuous Data
Gaussian Naive Bayes is a variant of Naïve Bayes that follows Gaussian normal distribution and supports Continuous data. The Gaussian or Normal distribution is the simplest to implement as the user is required to calculate the mean and standard deviation for the training data. It overall provides better performance by eliminating insignificant specifications. The algorithm uses probability for many classification functions while other functions are used to estimate data distribution.
II. RESEARCH METHODOLOGY
A survey was given to Goa students in various classrooms via online Google forms in order to get information on the learning curve of students during COVID-19. There were options to choose from when answering the survey's questions. The best choice that reflected their values had to be selected by the pupils. This allowed us to collect responses from around the State by disseminating the poll from September 30 to October 10, 2021. After reviewing the data, we chose approximately 461 records with the intention of obtaining a wide range of replies.
A. Data Collection
We gathered the information by sending out questionnaires via Google Forms to our Goa-based students, friends, family, and other well-wishers.
B. Data Representation
Email address, name, educational level, name of the institution, age, gender, taluka, and a few other columns have sub-questions make up the total of 36 columns, including the timestamp which is the default.
C. Data Preprocessing and Cleaning
Excel was utilised to pre-process the data, and a Jupyter notebook was used for analysis and analytics. You may create and share documents with live code, equations, visualisations, and text with this open source web application.
We used a variety of steps for pre-processing.
First, we converted any suitable string values to numeric numbers. The data was then filtered, and outliers were eliminated from the necessary rows. This decreased the number of rows from 561 to 423 rows. We then generated 5 graphs using the filtered data that were required for further investigation.
D. Data Analysis
After cleansing and preprocessing the data, feature selection was done. To achieve the best accuracy, we took 22 columns out of 35 columns. Then, 2 columns—Sum and Final Result—were added. The sum column contains the total count for each row that was calculated, and the final result column contains the average of all responses for each individual. All of this was done using simply Excel, where 0 means performance has not increased and 1 means performance has.
E. Data Analytics
To achieve the best accuracy, we used feature selection on our dataset and removed 22 of the dataset's 35 columns. The data was then divided into training and testing. The Random Forest Classifier technique was then used to create baseline models, and five other algorithms were employed to assess the accuracy.
III. RESULTS AND ANALYSIS
A. Data analysis
The pupils were given a multiple-choice grid with five distinct options for each statement (only one of these could be selected for each statement). Agree, Disagree, Neutral, Strongly Agree, and Strongly Disagree were the available responses for each statement. The statements were categorised into five main groups: the ability to adapt to online classes, problems with online classes, mental health problems caused by online classes, the effectiveness of the online teaching and learning process, and the overall effect on personality.
1) Adaptability To Online Classes
50% of students believe that e-learning tools are easy to use and 50% say they can do tasks faster in online mode. Most of the students strongly believe that online classes are not better than normal classroom classes. 70% agree that teachers are putting lots of efforts into making it easier for students to learn online
The accuracy of the GaussianNB() algorithms has increased when compared to the LinearDiscriminantAnalysis() algorithm, but it has decreased (by 95.29%) when compared to the RandomForestClassifier() algorithm, LogisticRegression() algorithm, K NeighborsClassifier() algorithm, and DecisionTreeClassifier() algorithm, all of which have accuracy scores of 100%..
TABLE I ALGORITHM ACCURACY SR.NO. ALGORITHMS ACCURACY 1 RandomForestClassifier() Algorithm 100% 2 LogisticRegression() Algorithm 100% 3 LinearDiscriminantAnalysis() Algorithm 94.11% 4 KNeighboursClassifier() Algorithm 100% 5 DecisionTreeClassifier() Algorithm 100% 6 GaussianNB() Algorithm 95.29% According to the preceding table, all four techniques, with the exception of LinearDiscriminantAnalysis() and GaussianNB(), provide 100% accuracy. 354 students\' grades have increased, whereas 69 students\' grades have not.
 Abdelsalam M.M., Ebitisam K.E., Shadi A., Hasan R. &Hadeel A. (2021). The Covid-19 Pandemic And E-Learning: Challenges AND OPPORTUNITIES from The Perspective Of Students And Instructors. Journal of Computing In Higher Education. Doi.org/10.1007/s12528-021-09274-2  Dr. Wahab Ali (2020). Online and remote learning in higher education institutes: A necessity in light of Covid-19 Pandemic. Higher Education Studies. Vol.10, No.3.  EdyBudiman. (2020). Mobile Data Usage On Online Learning During Covid-19 Pandemic In Higher Education. iJIM. Vol. 14. No. 19.  F. Zheng, N. Abbas Khan, S. Hussain. (2020). The Covid-19 Pandemic And Digital Higher Education: The Impact Of Students’ Proactive Personality On Social Capital Through Internet Self-Efficacy And Online Interaction Quality. Children And Youth Services Review. Doi:https://doi.org/10.1016/j.childyouth.2020.1055694  GhadaRefaat El Said. (2021). How Did Covid-19 Pandemic Affect Higher Education Learning Experience? An Empirical Investigation of Learners’ academic Performance at a University in a Developing Country. Advances in Human-Computer Interaction. Vol. 2021. ID6649524.  Haozhe J., Atiquil A.Y.M., Xiaoqing G. & Jonathan M.S.(2021). Online Learning Satisfaction in Higher Education during the Covid-19 Pandemic: A Regional Comparison between Eastern and Western Chinese Universities. Education and Information Technologies. https://doi.org/10.1007/s1063-021-10519-x.  Joana P., Ariadna L., Frances S., Marc A. & Daniel A. (2021). A Methodology to Study the University’s Online Teaching Activity from Virtual Platform Indicators: The Effect Of The Covid-19 Pandemic at UniversitatPolitecnia De Catalunya. Sustainability 2021, 13, 5177. https://doi.org/10.3390/su13095177  Lokanath M., Tushar G. &Abha S.(2020). Online teaching-learning in higher education during lockdown period of Covid-19 pandemic. International Journal of Educational Research Open. 2020. 100012  Maria J.S., Sandro S. (2020). The Covid-19 Pandemic as an Opportunity to Foster the Sustainable Development of Teaching In Higher Education. Sustainability 2020. 12. 8525; doi:10.3390/su12208525  Marion H., Melanie S., Michaela G., Barbel K., Svenja B., & Albert Z.(2020). Digital readiness and its effect on higher education students’ socio-emotional perceptions in the context of the Covid-19 pandemic. Journal of Research on Technology in education, DOI:10.1080/15391523.2020.1846147  Monika S., Ashish K., &Gursharan K. (2020). Research Perception, Motivation and Attitude among Undergraduate Students: A Factor Analysis Approach. Procedia Computer Science. Vol. 167. 185-192  N. Kapasia, P. Paul, A. Roy, J. Saha, A. Zaveri, R. Mallick, B. Barman, P. Das, P. Chouhan. (2020). Impact Of Lockdown On Learning Status Of Undergraduate And Postgraduate Students During Covid-19 Pandemic In West Bengal, India. Children and Youth Services Review. doi: https://doi.org/10.1016/j.childyouth.2020.105194.
Copyright © 2022 Nisha Sawant, Dnyandev Ravindra Khadapkar. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Authors : Nisha Sawant
Paper Id : IJRASET48311
Publish Date : 2022-12-22
ISSN : 2321-9653
Publisher Name : IJRASET
DOI Link : Click Here