Vegetable Price Prediction Based on PSO-BP Neural Network

Abstract

In order to predict vegetable price accurately, 117 sets of green pepper and related factors price data from 2012 to 2015 in Dan Zhou city were selected as the sample data, of which 100 groups were training data and 17 groups were test data. Based on analyzing fluctuant features of vegetable price, with the global stochastic optimization idea to optimize initial weights and thresholds of back propagation (BP) neural network, the PSO-BP prediction model concerning vegetable retail price was set up by using the particle swarm optimization (PSO) algorithm. The experimental results indicated that compared with the traditional BP method, the PSO-BP method could overcome the over-fitting problem and the local minima problem, effectively reduced training error and increased the predicting precision.