Sentiment analysis on twitter for predicting stock exchange movement

Abstract

This paper is proposed to build a model by applying two methods, namely support vector machine and nonnegative matrix factorization in the process of predicting stock market movement using twitter and historical data. The stock exchange is based on the LQ 45 stock with period from August 2018 - January 2019. The features consist of closing price, volume, percentage of topics and sentiment. The price and volume are taken from yahoo finance data, while topics and sentiment are taken from comments of each stock market in LQ45. NMF method is used to get the topic percentage feature in the tweet data, while the analysis sentiment value is obtained using SVM. The evaluation results obtained by getting the level of accuracy using confusion matrix. The accuracy value of stock movement predictions in this study is 60.16%.