Supporting Investment Decision Using Socio-Economic Issues Exploration and Stock Price Prediction

Abstract

Today, technology becomes the effective medium for spreading the information. They bring news and issues to reach a vast audience, including business-related news. The socio-economic implications are huge, one of them impacted the changes in stock prices and creates the value uncertainty. Therefore, the requirement of market information and stock prediction is crucial for investment decision making. the of this study is first to explore the relations between socio-economic issue and stock price, second to predict stock price future value. We choose IDX LQ 452017 as the case study. We use sentiment analysis to fulfill first objective and ANN model using their highest and lowest price, open and close price, and also transaction volume for second objective. We use Support Vector Machine (SVM) and Naive Bayes Classifier (NBC) as training classifier according to time and polarity. The results show 75% positive correlation between sentiment and the dynamics of stock prices. SVM classifier gives a better result than Naive Bayes Classifier, proved by 62% precision, 52.33% recall, 78.01% accuracy, 56.75% f-measure and 0.1 KAPPA. Our ANN based prediction model able to predict future stock price with 99.93% accuracy level. The conclusion of this research is sentiment analysis and stock price prediction have the potential for supporting investor decision using socio-economic issues and stock price prediction.