DEVELOPMENT OF AN AGENT-BASED FRAMEWORK FOR STOCK MARKET TRADING

Abstract

Online stock trading is a system that allows buying and selling of stocks through the Internet. Online trading demands time and bandwidth. Not only this, Entrepreneurs in stock trading need adequate knowledge of how stocks are being traded in stock exchange market in order to enhance their skills in stock trading thereby minimizing possible loss when they enter the real stock market. This project presents the development of an Agent-based Framework for stock trading. The system made use of software agents to carry out stock trading on behalf of traders. The components of the system include: Market maker agent, Informed agent, Uninformed agent, Information agent and the Hold agent. Each user creates any of these agents. Initial capital are allocated to the investor’s agents, the agents now use the capital to trade on behalf of the investors and making decision to meet the user’s investment goals which include optimum return on investment. The key issues addressed include assembly and incorporating diverse information sources with trading agents and providing decision-making for investor’s agent in the stock market using trading indicators for informed agent. A prototype of the system has been implemented in Java Application Development Environment (JADE) platform. Agents were created by the users. Initial capitals were allocated to the agents to trade with over a specified period of simulation. Their returns on investment were recorded. A comparative analysis of the strategies of stock trading by these agents during the trading period was carried out in terms of the returns on their investment. The result shows that different strategies yield different returns on investment. This work will be a useful tool to stock traders in learning the dynamics of the stock market and will serve as a training platform for learning stock trading. Future work will try to investigate the use of intelligent agents in stock trading.