Sentiment analysis on land and forest fire management in Twitter using Naïve Bayes method

Abstract

Forest fires that occur every year in Indonesia, especially on the islands of Sumatra and Kalimantan are a big threat to the environment. Furthermore, this disaster has caused huge losses not only to Indonesia but also to its closest neighbor. Public opinion about this is widely expressed on social media. Considering that Twitter is one of most well-known social media used, a study on sentiment analysis to find out which direction people tend to think about handling forest and land fires in Indonesia was then carried out. Public sentiment is categorized into both categories positive and negative sentiments. This study takes data from people's tweets about handling forest fires in Indonesia. The stages of the process are crawling, case folding, tokenizing, filtering, stemming, then Term Frequent-Inverse Document Frequent algorithm for document extraction, and comparison calculations using the Naive Bayes method. The algorithm of Naive Bayes was successfully applied to this study by producing a classification of 90% of Twitter users giving a negative response, while the remaining 10% gave a positive response to forest and land fires.