Thailand Stock Index Forecasting Based on Multiple Data Sources Using RNN-LSTM Approach

Abstract

Stock investment is an interesting option that has attracted the attention of investors. This research presents Thailand Stock Index Forecasting, a case study of the Stock Exchange of Thailand (SET) applying Recurrent Neural Networks with Long Short-Term Memory (RNN-LTSM). The research used multiple data sources 15 years of SET Market Data from 2006 to 2020 for stock forecasting divided into Training Set 80% and Test Set 20%. Mean Square Error (MSE) was used for the evaluation model. The results showed that RNN-LTSM Approach optimized the model’s performance based on 5 stock asset companies used in this research. Therefore, the proposed model can be used to develop related forecasting or recommendation systems in the future to help investors consider the risks of investment and support investment decisions.