Opinions or sentiments given by the public on Twitter social media have attracted attention for research. In this study, what will be done in connection with sentiment analysis on Twitter is to use a target-dependent approach. Where given a data twit or opinion relating to 3 major cities in Indonesia, namely Jakarta, Bandung and Medan, and will be classified as opinions that are positive, negative or neutral. Due to the social media used is Twitter, which can only write opinions in 140 characters, it is possible to find data on tweets or ambiguous opinions. Therefore, this study focuses on sentiment analysis with a target-dependent approach using the SVM method for classification. The target-dependent approach itself consists of 2 stages, namely determining matters relating to the topic of discussion and collecting data on tweets related to the topic. Broadly speaking, this research was carried out in 4 stages, namely preprocessing, classification with a target-dependent approach and SVM method and finally the determination of opinion classification)
Copyrights © 2020