基于新疆粮食产量的灰色关联分析及其BP模型预测

为明确影响新疆粮食产量的主要因素及预测未来变化，采用灰色关联法和BP神经网络预测模型，对2000—2019年影响新疆粮食产量的9个关联指标进行分析。结果表明，粮食作物播种面积、劳动力数量和有效灌溉面积是影响新疆粮食产量的主要因素，其关联度均高于0.91。从新疆的实际情况和关联度分析出发，确定影响粮食产量的6个重要因素是粮食作物播种面积、就业人数、有效灌溉面积、农业机械总动力、化肥施用量和新疆人口数量。利用matlab2015b软件构建BP神经网络模型，预测2020年新疆粮食产量为1 542.7万t，预测值与当年的实际粮食产量相差不大，说明BP神经网络模型对粮食产量的预测具有很好的匹配性。

Study on the Grain Yield in Xinjiang Using Grey Correlation Analysis and BP Model Prediction

To determine factors affecting the grain yiled in Xinjiang and to predict its potential changes in future, the BP neural network prediction model and grey correlation method were used in this study to analyze nine ralated indices from 2000 to 2019 on grain yield in Xinjiang. The results showed that sown area of grain crops, agricultural labor quantity and the effective irrigated area were the main factors affecting the grain yield in Xinjiang, their correlation factors were all higher than 0.91. Based on the actual situation and correlation degree analysis of Xinjiang grain production, six important factors affecting grain yield were determined i.e., sown area of grain crops, agricultural labor quantity, effective irrigated area, total power of agricultural machinery, fertilizer application and population of Xinjiang. Matlab 2015b software was used to build BP neural network model which predicted that the grain output in 2020 for Xinjiang was 15.427 million tons, the predicted value was not much different from the actual grain output of that year, which showed that the BP neural network model had a very good prediction over grain output.