| Paper is divided into five parts, the first part of the background of the investment fund and investment fund research at home and abroad with a brief introduction, the second part of the theory of investment funds were introduced, including portfolio theory and Markowitz capital asset pricing model, the third part of the Fund's investment rating system was introduced, the fourth part of the Fund the Fund to evaluate the benefits and the wind direction, the fifth and sixth part focuses on some historical data and investment risk preferences Construction of composite portfolio will be the seventh part is the simulation of future market conditions, combined with investors prefer to build portfolio.In the third section, we use factor analysis, through the respective benefits and risks of the fund's representation of the two synthetic indicators, are integrated, comprehensive indicators to overcome the shortcomings of single indicator and indicators for duplication of information between. We use foreign Funds Performance Evaluation model commonly used in CAPM, Jensen indices and risk rate of return and other indicators of 192 funds November 3, 2008 to April 30,2009 income, to evaluate the risk profile. Then based on the evaluation results,192 funds according to their risk and return characteristics into nine categories, the formation of a comprehensive evaluation and improvement of the model that direction.By calculating the value of a specific number of portfolio risk to measure changes in the risk of law, we have established a portfolio risk model, respectively,192 calculate the risk of random combinations of the Fund, and the classification based on the classification of the nine random portfolio risk, and simulated by Matlab combination of the two cases, the risk Biandong map, compare Dechu random combination of risk-classification was Hao risks in all random combinations, thus confirming the importance of this classification, while the following portfolio analysis basis.We return from the fund portfolio and risk modeling in consideration, the optimal portfolio problem as a multi-objective programming problem. Meanwhile, the introduction of risk preference coefficient as an indicator to investors to invest in the style of multi-objective programming model with parameters into a single objective nonlinear programming problem. In the process of solving the problem, due to non-short conditions, first of all solve the problem of unlimited planning solution can be short, and two programming problems using the relationship between the optimal solution (the relations are detailed in the model Explanation and demonstration) is not short of plans to further search for the optimal solution. At the same time, we have established in the MATLAB environment using genetic algorithm to directly solve the planning problems of the model, finally, we use the second part of the model is the result of the Fund's classification rating out of 10 funds selected as model validation, and compared and analyzed.In the fourth section, we use the funds based on RBF neural network prediction, the neural network approach any nonlinear continuous function with the ability to establish a relationship between input and output to predict the future trend of the Fund. From the Matlab simulation results, the model price for the Fund's short-term prediction can achieve better results. |
PDF Full Text Request