Application Of Data Mining Technology In Shanghai Stock Index Forecast

As China’s reform and opening up drive deepens, construction of financial market gradually improved, financial market gradually focused, the Chinese Stock market was chased after from 1990. The stock market was the focus spot because of higher income and risk. In China, investor sentiment has been improving gradually in recent year. In 2004, the number of opening the stock account was creates new high. With high risk and high earning, how to get higher income and lower risk, that we focus on. The stock price forest correctly, according to the history record, we can forest the future stock price, and improve the reliability information source. Now, by data mining technology and machine learning, we improve a future stock price to investor.Because the stock price is floated by demand. The demand is floated by some active and passive reason. Investor ahead of time to forest the stock price. We can provide a gist to investor. According to the gist, investor layout investment program. It’s can improve the level of resource utilization. All above base on the precisely forest the stock price.Related literature appear, wavelet neural network forecasting the time series data have good effect. But this paper find that using wavelet neural network forecasting spend much time for training sample. On the contrary, in some special time quantum wavelet neural network forecasting value is much more limited in true value than SVM. In this paper, we use SVM to forecast the same things. And finally, this paper compare the SVM with two kinds of neural network.This paper base on Shanghai composite index, forecasted empirical analysis, compare the SVM, BP neural network and wavelet neural network. And compare high-speed short term passenger flow forecasting, MSE, computing time and etc. Wavelet neural network base on BP neural network to develop. And Wavelet neural network is derive by BP neural network. Wavelet neural network computational efficiency is lower than SVM. In this empirical analysis, the accuracy of wavelet neural network is better than accuracy of BP neural network. SVM is one of the best arithmetic in this three arithmetic. Whatever computational efficiency and accuracy were well.