Stock Market Prediction: The Application of Text-Mining in Vietnam

There are very few studies in Vietnam on the application of text mining in finance and Vietnamese language processing. The origin of this study comes from one of the leading studies on the use of machine learning to analyze text data from 4 well-known online newspapers in Vietnam to forecast the increase, decrease and neutrality of the VN-Index one day in advance. This study used nearly 70,000 articles from four reputable and reliable online newspapers in Vietnam as input data for machine learning models. These were: decision trees, random forests, KNNs and SVMs. After selecting the best model (SVM) and the best dataset (Vietstock), the techniques used to dig deep and refine the findings raised the accuracy to 60.1%. The end result is solid evidence that news about the financial and stock situation in the popular press affects the price movements of the VN-INDEX and the Vietnamese stock market.