Computer market becomes more sophisticated and price forecasting is gaining importance for market participants to adjust their market behavior. In the past, price forecasting of computer market is mainly concentrated in the forecasting of the computer itself such as desktops, laptops and so on. Whereas a contrast resolution is proposed to deal with this issue that forecasts price trends of computer market through price forecasting of main computer accessories which can make the forecasting window in advance. Two classic model, Back Propagation (BP) neural network and Support Vector Machine (SVM), are introduced to implement the week-ahead price forecasting of computer accessories. The simulation results show that SVM model is better than BP neural network model for its higher forecasting accuracy. Under the same forecasting conditions, the Relative Errors of SVM model is 1.87% lower than that of BP NN model and the mean absolute errors of SVM model is 17.91% lower than that of BP NN model. Therefore, price forecasting for computer accessories based on SVM is valuable and feasible for computer market which can provide richer and more accurate analysis information of price trends for market participants in advance, with a high reference value.
展开▼
