Bull and bear : a personal investment assistant

Abstract

Finance is one of the many areas where technological developments have led to a paradigm shift in the way the industry operates at all levels. This includes the area of financial markets and investments. Today, markets are accessible to sophisticated and non-sophisticated investors alike. In addition, investors today have access to immeasurable amounts aggregated qualitative and quantitative information that is made available by news information providers, technology or private companies that provide freely accessible historical archives, organisations that process and distribute news and Government agencies. Based on the availability, accuracy, timeliness and accessibility of such information, this study attempts to apply artificial intelligence in the investment decision making process. The ultimate aim is to be able to provide an investor with an investment suggestion, for example, whether to buy or otherwise, a specific share. The adopted approach is that of integrating both quantitative information (share prices, trading volumes and other metrics) as well as qualitative information (company specific news) in the investment suggestion system. This is done through the combination of a purposely build neural network and through company specific news sentiment analysis. With respect to the neural network, a back propagation algorithm is used to experimentally select the most appropriate parameters for forecasting time series. Sentiment analysis score values are used to quantify the sentiment of specific company specific news. The object of the system design and implementation are to studying the possibility of accurate share price movement predictions though the adopted approach. The assessment of the quality of results is performed through statistical evaluation as well as through the post-occurrence examination of results and investment suggestion for the forecasting time window.