Mining textual terms for market prediction analysis using financial news

Stock market is uncertain and it is not entirely predictable. Financial investors use fundamental analysis and technical analysis to study the movement of stock prices. Fundamental analysis makes use of quantitative and qualitative factors such as financial news to drive investment decision. But, it is tedious to digest textual articles from numerous news sources. Besides, Bursa Malaysia sees little studies on the use of algorithms to predict stock prices. This study focuses on the use of machine learning algorithms to construct a model that is able to predict Bursa Malaysia stock prices. In this research, we concentrate to identify linguistics terms from financial news that can contribute to the movement of stock prices and to develop a prototype that can classify sentiment of financial news for investment decision. We experimented with five companies from different industries of the top market constituents in FBMKLCI. First, financial news of these companies from The Edge Markets amounted to 14,999 articles were crawled to be used as the dataset. In addition, the dataset was processed and transformed into TF-IDF representation. S VM was used to train and test the dataset, and the accuracy recorded was measured to be at 56%. Hopefully, the findings of this research can be used to assist investors in investment decision making.