Stock Prediction System Based On Neural Network Design And Implementation

Now, when the stock market in the socialist economic construction in China plays an increasingly important role, stock investment has gradually become an important investment. Under the condition of this reality, the research and development on intelligent stock picking and stock price forecasting system, is full of important social and economic significance.System downloads the historical transaction data and the real-timetransaction data of stocks, calculates the value of the various technical indicators, and then write to the database, studies and summarizes several strategies for selecting stock according to morphological features of technical indictors, selects the stocks of meeting the conditions of technical indictors querying stocks’ indictors in database, and predicts the stocks short-term movement and prices according to their technical pressure and support price levels. In stock forecast, stocks’ K-line chart and various technical indicators charts are drawn, by adjusting the price of the stock we can emulate the market transactions of the next day. At this time, observing the real-time changes of stock’s K-line chart and various technical indicators charts, we can measure the general trend and the stock price of the next day.A large number of simulation results show us that in stock price (index) system neural network provide colossal application space. Although we have made a little progress in predicting stock market based on neural network, which is just tentative and elementary, we need further research.The thesis integrates Delphi which is programming flexibly and efficiently with Matlab which has powerful computing ability. The system’s main features include steady and rapid running speed, easy manipulation and powerful functions.