A Decision Level Identity Fusion Based Prediction Model for Smooth Fluctuation Products

Nowadays, various indicators with different features are used to analyze the price trend and forecast the future prices for investment decisions. However, for investment products that do not have drastic price trend changes, such as gold, when making decisions on whether to buy or sell based on the predicted price usually does not have good returns. Therefore, the prediction model for investment products with smooth fluctuation (PIPSF) is proposed based on the idea of decision level identity fusion. In this model the 'ideal buy and sell points' is established to decide the buying or selling of investment products. In the case of the large amount of data and feature vectors, the elastic net regression was used for feature selection, it combines the advantages of two regularization methods to ensure both the goodness of fit and the generalization ability of the model. The experimental results show that the model proposed in this paper has higher returns when compared with direct price prediction.