Design And Implementation Of Stock Buying Point Forecast System Based On Neural Network Technology

Stock market is affected by political situation, national policy, investor psychology and other factors. Its interior law is very complex, and change cycle is unordered. It’s difficult for people obtain hidden information from enormous data by intuitively and experience. It is a necessary choice to introduce data mining technology to analyze and predict stock market. Represented by neural network technology, it is the human brain’s simplification and abstraction and mainly depends on intelligent decisions and deduction. Since the neural network has the strong characteristics of self-organization, self-adaptation, self-learning and so on, it can be distributed and processed, which has obvious advantages in the application of the complex and changeable nonlinear dynamics system of the stock market.This paper designs and implements the stock prediction model based on neural network technology. Compared to similar studies, this prediction model innovative cumulative increase to the maximum period of time in the future as a prediction target.It can be more accurate to predict the ups and downs of the stock trend in average accuracy rate training set and test sample set of more than 80 percent, with strong practical application value. This model consists of six independent BP neural network composed by six groups feature information respectively training and learning, to realize the stock ups and downs of the next 28 days forecast.This prediction model design and implementation of the stock buy-point prediction system based on.This system is divided into client subsystem and intelligent analysis subsystem. The client subsystem includes data volume loading module and the predictive query module. The intelligent analysis subsystem includes data acquisition module, feature extraction module and training study module.Through the feature extraction of daily data of Shanghai and Shenzhen’s more than 2000 stocks between January, 2005 and January, 2016, using neural network for training study and predicting stock’s buy point offer important references for stock investors when trading stocks.