A Method for Stock Trading Strategy Combining Technical Analysis and Particle Swarm Optimization

Accurate stock trend prediction is a difficult job because various intricate and complex factors affect changes in price, trading volume and trends of a stock market. On a macro scale, the factors could be the overall global economic environment, industry trends, individual economic environment (business operation and competitors’ development), the amount of floating capital in the market, etc. Likewise, on a micro scale, the factors could be investors’ decision-making behaviors, their psychology, the weather, and so on. How to apply these real trading information from the stock market to establish and construct an accurate forecasting model to identify correct trading signals and noises is a mammoth challenge. This research utilizes the Particle Swarm Optimization (PSO) Method and combines the Technical Analysis tools Simple Moving Average(SMA) and the principles of crossover trading decisions to establish a stock trend prediction model and a trading decision model. The target identified for this research is Taiwan Stock Market and the prediction model is compared with the buyand-hold strategy. The experimental results indicate that most of the stocks in the forecasting model strategy perform better than in the buy-and-hold strategy and thus, this forecasting model strategy is a valid investment strategy.