Commodity Futures Price Prediction and Trading Strategies -- A Signal Noise Difference Approach

This paper introduces the signal noise difference method and applies this method into the commodity futures price prediction. Based on the prediction rules mined from the data of 25 potential prediction indicators of SHFE CU, a corresponding transaction strategy is established. And we use the market data from 2009 to 2013 to test our transaction strategy, which obtains a result of 147.85% annual yield. In addition, several improvements are discussed to optimize this model.