REVIEW SENTIMEN ANALISIS APLIKASI SOSIAL MEDIA DI GOOGLE PLAYSTORE MENGGUNAKAN METODE LOGISTIC REGRESSION

Abstract

Everyone has their own nature and way of thinking, which makes their style of conveying what they are thinking is different from one to another. Not only that, what they are conveying can have a different meanings according to the person who listens to it and which perspective they take it into. From this, the reviews given have different meanings for each person which other people can easily understand the meaning behind it. Although humans can distinguish whether the review given by other people has a positive or negative meaning, machine cannot understand this without being given instructions first, in order for machine to find out the meaning in the text, Therefore, this research was conducted. The purpose of designing this system is to make it easier to analyze a collection of reviews on a social media application from google playstore without the need to see all the review given by other users. This system is also designed to fulfill the purpose of evaluating using the TF-IDF method and the Logistic Regression method on the program to be created. The intended evaluation is from the level of accuracy obtained from the use of the model formed using the method chosen and how many predictions have the correct value based on the result of Confusion Matrix from the program that has been designed . From the results obtained, it can be concluded whether the social media application has a good reputation from its users or not.