基于BP、LSTM和ARIMA模型的蔬菜价格预测

Vegetable price prediction based on BP,LSTM and ARIMA models

为系统统计蔬菜价格,实现蔬菜价格可视化并加以预测,以利于生产者科学决策。为此,首先爬取广州江南果菜批发市场所有的蔬菜价格,并对蔬菜价格的数据集进行预处理,然后建立起基于时间序列的ARIMA预测模型、BP神经网络预测模型和LSTM神经网络预测模型,通过3种模型对爬取的蔬菜价格进行分析和预测,最后将3种预测模型的实验结果进行对比。在选取的多种蔬菜的预测结果中,LSTM、BP、ARIMA模型的相对误差小于1%的平均比例分别为0.037、0.07、0.097,相对误差小于5%的平均比例分别为0.215、0.338、0.433,相对误差小于10%的平均比例分别为0.436、0.573、0.694。结果表明,ARIMA模型在预测蔬菜价格方面的准确率比LSTM、BP模型更高。

The main purpose of this study is to systematically calculate vegetable prices and visualize and predict vegetable prices so as to facilitate the scientific decision-making of producers.Therefore,it was used Python Programming Language and first crawled all vegetable prices in Guangzhou Jiangnan Fruit and Vegetable Wholesale Market and pretreated the data sets of vegetable prices.Then it was established ARIMA prediction model based on time series,BP neural network prediction model and LSTM neural network prediction model.Through three models,the crawled vegetable prices were analyzed and predicted.Among the prediction results of the selected vegetables,the average proportion of the relative errors of LSTM,BP and ARIMA models less than 1%was 0.037,0.07 and 0.097respectively,the average proportion of the relative errors less than 5%was 0.215,0.338and 0.433respectively,and the average proportion of the relative errors less than 10%was 0.436,0.573and 0.694respectively.Finally,results of the comparion between three models showed that ARIMA prediction model was more accurate in predicting vegetable prices.