Soybean futures price trend is affected by many factors being nonlinear. In order to improve the accuracy of soybean futures price forecast, a soybean futures price forecasting model based on combination of pursuit algorithm and recurrent neural network is proposed. Choose soybean futures data from April to August 2019, using the projection pursuit algorithm to select the optimal index and predict it as the input of the recurrent neural network. The  analysis results show that the combined model has good predictive performance which can provide some reference for investors short-term forecast.

projection pursuit, recurrent neural network, soybean futures, price forecast.