The Effect of Chi Square Feature Selection on the Naïve Bayes Algorithm on the Analysis of Indonesian Society's Sentiment About Face-to-Face Learning During the Covid-19 Pandemic
Ayunda Firdaus Chusnul Habiba, R. Rizal Isnanto, J.Endro Suseno · JST (Jurnal Sains dan Teknologi) · 2023
Criticisms and comments submitted to the Indonesian people regarding government policies regarding face-to-face learning during the Covid-19 pandemic have pros and cons. Not a few parents are worried about this policy because parents are still afraid of the spread of the new cluster of Covid-19 in Indonesia which is growing while on the other hand many people who think this is good to apply considering that online learning is considered less effective because many students are difficult to receive material delivered by teachers online and many students who do not have adequate tools. The number of opinions written on Twitter requires classifying according to the sentiments they must get the tendency of these opinions whether they tend to have neutral opinions, positive or negative. Sentiment analysis in this study was conducted using the Naïve Bayes method and chi square's characteristic selection in conducting classification. The results of the analysis of the Naïve Bayes method with chi square characteristic selection have an accuracy of 93% and without chi square characteristic selection has an accuracy of 92%, so it can be concluded the Naïve Bayes Method with chi square characteristic selection has a better accuracy compared without using chi square characteristic selection.