Research Of Continuous Maximum Entropy Stock Price Prediction Model

Under the increasingly economic situations and sophisticated political of domestic and international environment, the financial market will usher in the great opportunities and challenges. In the environmental context, The stock investors how to accurate forecasting the future stock price, and get close to the expected return, has become one of the important issues. So, through effective way to successfully predict the stock has gradually become one of the top ten issues facing the academic stage. In China, stock investors calculating the rate of return often used percentage method, because this formula is easy to accept for the primary investors, and many books, newspapers and magazines are referenced percentage method. So we based on the maximum entropy models for J.M.Cozzolino. Changes the formula of rate of return, to construct a new utility function, the utility function can reflect the real expectations of investors, it substituted into the maximum entropy models, so the model is transformed into a stock forecasting model into a new model, namely continuous maximum entropy stock price forecasting model. The objective function is the stock price as a continuous random variable maximum entropy function.so the utility function constraints are a prior conditions need to be met. In the process of solving, using the transformation of random variables, introduces the Lagrangian function.at last we can obtain continuous maximum entropy stock price forecasting models. As in the real stock trading will generate transaction costs, the paper adds to the transaction costs the utility function makes the model more realistic,which can more accurately predict future stock market stock prices. So this thesis does a research on the maximum entropy and contains five chapters.The first chapter describes the background and significance of the entropy optimization theory, and describes literature reviews, including stock prediction and entropy function. Finally introduces the innovations and the main content of this thesis.The second chapter describes the generation of development and evolution of the entropy, following describes the basics knowledge of the entropy. Finally, lists a series of entropy in the investment in the application.The third chapter describes the utility function and the principle of maximum entropy, and later introduces the continuous maximum entropy model, then we can change the rate of return to change the utility function, so we can optimization of the continuous maximum entropy model. Then we can creating a continuous maximum entropy stock price prediction model, in the final, the new model is verified by example.The fourth chapter is mainly based on the basis of chapter 3, and then adding the transaction costs to the new model. At last, the new model is verified by example.In the last chapter summarizes the main content and the main outcomes of the paper, and lists some problems of this paper, finally this master thesis ends.