A Mixed Model based on Wavelet Transform and Support Vector Regression to Forecast Stock Price

The prediction of financial price has always been an important reference indicator for investors to make investment decisions. This study uses the differential method to highlight the trend of the stock price change. To further suppress the influence of stock noise data, the wavelet transform is introduced to decompose the stock data into principal component and detailed component. Moreover, the sliding window average is introduced to suppress the fluctuation of stock price caused by noise. Stock price forecasting is performed by the Support Vector Regression (SVR) method. We chose 12 stocks from the US stock market and conduct an experiment. The result shows that the model proposed in this study can achieve a lower error on multiple criterions than the existing model which uses combination with sliding window, Firefly algorithm (FA) and SVR.