An agent based stock price prediction model for the Nairobi securities exchange

Abstract

The stock market is a key market in any economy and financial forecast such as stock price prediction is a field receiving much attention both for research studies and commercial applications. Stock market forecasters are keen on developing a successful approach to predict stock prices even more accurately since there is the motivation of gaining massive profits from trading shares by using well defined attractive strategies. This research project develops a stock price prediction model build using JADE based on multi agent architecture in order to harness the power of agents and provide investors with predicted trend of share price by incorporating the various correlated factors like economic, political, company outlook to traditional price over time, demand and supply in order to accurately forecast the stock price trend and thus provide a buying or selling signal to traders. The Trend is determined by incorporating text processing in the agents from live news sources. The model was tested and proved a key tool for stockbrokers, novice traders and investment bankers sinceits automated and more robust than the traditional methods of price prediction. Key Words; Agents, Stock Market, Prediction. Stock price. Model. Text processing. JADE.