SENTIMENT ANALYSIS OF PEOPLE DURING COVID- 19 USING SVM AND LOGISTIC REGRESSION ANALYSIS
Sayan Majumder, Anuran Aich, Satrajit Das
Assistant Professor, Gargi Memorial Institute of Technology
Starting from China, novel coronavirus is now the most dangerous threat to humans, all over the world. India is also not an exception. Besides using mask and sanitizers, Indian Government has also decided to maintain proper social distancing or lockdown. Our sentiment during this lockdown period also varied from man to man. In this paper, we have collected twitter data of people across India, during March to June and then using NLP, the polarity is measured, i.e. positive, negative or neutral. After that. We have used SVM classifier and Logistic Regression analysis to classify the sentiments of people. Python programing language is used in Anaconda distribution to simulate the result. After simulation, we have found that SVM gives 91.50% accuracy, whereas Logistic Regression gives 87.75% accuracy. At last, a comparative study between these two results, is also represented.
Keywords: Sentiment Analysis, NLP, Machine Learning, SVM, Regression.
Journal Name :
EPRA International Journal of Multidisciplinary Research (IJMR)
Published on : 2022-06-04