Trading Decision Maker: stock trading decision by price series smoothing and tendency transition inference

Financial engineering, such as trading decision-making, is a major topic for research and also has commercial applications. The stock price series has a series of change-points, and accurate prediction of its movements is the key to successful trading. However, making correct trading decisions is difficult because of the influence of embedded noise and price fluctuations that confuse the interpretation of stock trends. This paper proposes a novel stock trading method, called the Trading Decision Maker (TDM), based on price series smoothing to reveal important change-points, which reflect changes in stock trends more precisely. Tendency transition inference is used to identify these important change-points effectively. We demonstrate the usefulness of TDM in evaluating profitability capacity, and prove that its accumulated rate of return performed 199.05% better than other buy-and-hold strategies used by open-ended mutual funds on the Taiwan Stock Exchange Capitalization Weighted Index (TAIEX) from 2001 to 2003.