Leveraging fundamental, technical, and sentiment analyses with quantitative techniques and machine learning in trading

Abstract

Stock markets are likened to random walks due to their complex, dynamic and chaotic nature. They are influenced by a wide range of factors, including economic, political, psychological, and company-specific variables. These make stock market forecasting a knotty challenging task. This paper introduces ten fundamental indicators derived from financial statement reports, alongside a technical analyses indicators encompassing of 25 candlestick patterns, 8 chart patterns, and 21 technical indicators. The investigation applies these indicators to 30 companies listed in the Straits Times Index (STI) and the Dow Jones Index (DJIA). Findings reveal that both fundamental and technical indicators hold significant predictive power in the context of the STI, with their effectiveness exhibiting variation across different industries. These indicators, under the same parameters, do not demonstrate the same level of predictive ability when applied to the dataset in the DJI. Further exploration is conducted on a single stock, combining sentiment analysis with fundamental and technical indicators. This comprehensive approach seeks to provide a holistic nature of stock price movements and the potential of integrating diverse methods to enhance the accuracy of stock market forecasts.