A new method of financial time series prediction

To solve the defects of traditional neural network, a financial time series prediction system based on rough neural network is proposed. Firstly, rough set is applied to reduce the data of financial time series sample so as to remove the disturbance of redundant attributes, which overcomes the impaction of unrelated data that imposed on the performance of network learning and simplifies network structure. Secondly, by using rough neurons instead of the traditional neurons, the performance of network is improved, and the scope of the application of network is expanded. The test verifies the effectiveness of this method.