**Abstract**

Instagram is one of the world's top ten most popular social networks. Instagram is the most popular social networking platform in the United States, India, and Brazil, with over 1 billion monthly active users. Each of these countries has more than 91 million Instagram users. The number of Instagram users shows the various reasons and goals for them to play this social media. Social Media Marketing does not escape being one of the purposes of using Instagram, with benefits to place a market for their products. Using text classification to categorize Instagram captions into organized groups, namely fashion, food & beverage, technology, health & beauty, lifestyle & travel, this paper is expected to help people know the current trends on Instagram. The Support Vector Machine algorithm in this research is used in 66171 post captions to classify trending on Instagram. The TF-IDF (Term Frequency times Inverse Document Frequency) method and percentage variations were used for data separation in this study. This study result indicates that the use of SVM with a percentage ratio 70% of dataset for training and 30% of dataset for testing produces a higher level of accuracy compared to the others.

**Keywords:** Instagram, Support Vector Machine, Text Classification, TFIDF, Social Media
