ANALISIS SENTIMEN MASYARAKAT TERHADAP PAYLATER MENGGUNAKAN METODE NAIVE BAYES CLASSIFIER

Abstract

Online shopping is so popular with the public because it is easy and convenient to do. The convenience of online shopping is supported by the payment method via paylater. However, paylater also results in bad behavior such as impulse buying. Various responses from the community made researchers conduct research to find out the public's view of paylater. In this study, researchers tried to do sentiment analysis using the Naive Bayes Classifier and TextBlob methods from the TetxBlob library with the Python programming language. From the dataset collected via Twitter, it produces 405 data. Sentiment analysis using the Naive Bayes Classifier method produces a negative sentiment of 70.62% or 286 data, positive sentiment is 22.72% or 92 data, neutral sentiment is 6.67% or 27 data. Meanwhile, using the TextBlob method also produced more negative sentiment, namely 55.8% or 226 data, positive sentiment collected 33.09% or 134 data, neutral sentiment amounted to 11.11% or 45 data. Thus, it can be concluded that the community feels unfavorable towards the use of paylater. In testing the model with the confussion matrix, it can be seen that the Naive Bayes Classifier algorithm is more accurate by 91% compared to TextBlob which is only 61%.