Implementasi Algoritma Naive Bayes Multimonial dalam Menganalisa Sentimen Masyarakat di Twitter (Studi Kasus : Perpindahan Ibu Kota)

Abstract

The move to Indonesia's capital city received different responses from the public, pros and cons. This has an influence on the development of the capital itself. The impact of the move will be felt by the community, especially the people of the Kalimantan area or private developers who want to invest. This research will discuss the classification of public opinions regarding the transfer of the capital which are categorized into 3 categories, namely neutral, negative and positive. The aim of this research is to obtain the results of sentiment analysis or public responses to the move of the nation's capital. The benefit for the government or development is as input regarding the direction of development, the economy and policies regarding the development of the nation's capital city. Data collection was obtained from Twitter using scrapping techniques. The sentiment analysis stage is carried out by preprocessing first, namely by labeling and grouping. At the analysis stage, the Multinomial Naïve Bayes algorithm was used. This method was chosen as an analysis method because this method is able to measure the frequency of occurrence of words in documents more effectively. The classification results obtained from the 1049 data used were that 25.54% stated that public sentiment was neutral, 35.08% stated that public sentiment was negative, and 39.37% stated that public sentiment was positive. The accuracy rate was also obtained at 92.67%.