Relevance Classification of Trending Topic and Twitter Content Using Support Vector Machine

Twitter is one of the most influential social media recently and has become an important information portal. Users can use the trending topics feature or hashtags popular to see current popular news. However, trending topics are often irrelevant to the content being discussed. Many users take advantage of trending topics so that their tweet content gets more public attention. Therefore, a text classification system is needed to determine the relevance of trending topics to Twitter content. Relevance Classification using the Support Vector Machine method and related to politics topic that use the Indonesian language. The research steps were data collection, data labeling, text preprocessing, TF-IDF weighting, SVM classification, and evaluation and analysis. The results of classification results show an average accuracy value of 86%. The average of precision, recall, and F1-measure values are 75%, 72%, and 70%. In this paper, the Support Vector Machine method has good performance for relevance classification on Twitter content.