Research On Security Investment Decision Based On Data Mining Technology

On the basis of previous studies, this paper researches the securities investment problems from three main aspects, stocks selection decision, securities portfolio investment decisions and securities investment fund securities investment decisions. Data mining techniques are used to work out the following conclusion.(1) The rough set theory is used to select suitable stocks for value investment practice. The intrinsic relationship between the degree of each financial indicator and securitiesis calculated based on historical data, and multi-attribute decision method is used witha comprehensive subjective and objective weights to decide the optional stocks. Empirical studies have demonstrated good accuracy and scientific value of the investment stock picking method based on rough set theory.(2) Taking full account of the differences in different industry stocks, this paper selects five stocks from different industries to construct the portfolio.Due to the different levels of risk and return preferences for the investors, a new portfolio objective function is constructed based on the investors utility function, so that the model can be applied to different types of investors. Particle swarm optimization algorithm is choosed to solve the problem. By the way of constantly updating the velocity and positionby particle swarm generation, the optimal weight ratio is concluded to build a portfolio for investors.(3) The performance of China’s securities investment funds is evaluated from two espects, the ability to select stocks and choose opportunities, the profitability of the funds in both bull and bear markets. The evaluation results are expected to guide investors to choose suitable securities investment funds according to their income requirements and their risk-taking levels. By comparing various models, the conclusion are as followed:most fund managers have the ability to share positive stock selection and negative selection timing ability; and, most funds could gain a higher income than the market average in the bull market, but less in the bear market.