基于深度信念网络的粮食产量预测
|
摘要
为了研究地区粮食产量的预测问题,提出了一种基于深度信念网络的粮食产量预测模型,利用改进K-means算法构建数据集预处理模型,通过多层受限玻尔兹曼机构建深度信念网络的预测模型,并利用对比散度算法训练预测模型。以河南省西华县1996—2016年小麦产量、种植面积与降雨量数据作为应用研究实例,将1996—2013年的小麦产量、种植面积与降雨量数据作为建模样本、2014—2016年的相关数据作为测试样本,进行预测模型的研究。结果表明,基于深度信念网络的粮食产量预测模型的平均预测精度超过97%,说明深度信念网络适用于地区粮食产量的预测,为粮食产量预测提供了一种新方法。
In order to study the prediction of grain yield in a region, a prediction model of grain yield based on deep belief network is proposed. Using K-means algorithm to preprocess data sets, the network model is built by using the restricted Boltzmann machine, and the model is trained by the contrast divergence algorithm. Based on the wheat yield data, planting area data and rainfall data from 1996 to 2016 in Xihua county of Henan province is used as an example of application research. Among them, the data from 1996 to 2013 were used as model samples, and the data from 2014 to 2016 as test samples. The results show that the accuracy of wheat yield prediction model based on deep belief network is high, average prediction accuracy is over 97%. It is indicated that deep learning is applicable to the prediction of regional grain yield. It provides a new method for predicting grain yield.
In order to study the prediction of grain yield in a region, a prediction model of grain yield based on deep belief network is proposed. Using K-means algorithm to preprocess data sets, the network model is built by using the restricted Boltzmann machine, and the model is trained by the contrast divergence algorithm. Based on the wheat yield data, planting area data and rainfall data from 1996 to 2016 in Xihua county of Henan province is used as an example of application research. Among them, the data from 1996 to 2013 were used as model samples, and the data from 2014 to 2016 as test samples. The results show that the accuracy of wheat yield prediction model based on deep belief network is high, average prediction accuracy is over 97%. It is indicated that deep learning is applicable to the prediction of regional grain yield. It provides a new method for predicting grain yield.
引文
[1] 程伟,刘国璧.灰色系统理论在粮食产量预测中的应用[J].湖南工程学院学报(自然科学版),2008,18(3):64-67.
[2] 王惠婷.基于混合时间序列模型的粮食产量预测[J].统计与决策,2013(12):23-25.
[3] 闫福春,陈青,徐秀珍.影响小麦产量因素对产量形成的贡献[J].江苏农业科学,2012,40(10):78-80.
[4] 王柳,熊伟,温小乐,等.温度降水等气候因子变化对中国玉米产量的影响[J].农业工程学报,2014,30(21):138-146.
[5] 王莹,樊鑫,李豪杰,等.基于深度网络的多形态人脸识别[J].计算机科学,2015,42(9):61-65.
[6] 夏春江,王培良,张媛.基于深度信念网络的木材含水率预测[J].杭州电子科技大学学报(自然科学版),2015(1): 31-35.
[7] 杨旭华,钟楠祎.基于深度信念网络的医院门诊量预测[J].计算机科学,2016,43(S2):26-30.
[8] 郑志蕴,李步源,李伦,等.基于云计算的受限玻尔兹曼机推荐算法研究[J].计算机科学,2013,40(12):259-263.
[2] 王惠婷.基于混合时间序列模型的粮食产量预测[J].统计与决策,2013(12):23-25.
[3] 闫福春,陈青,徐秀珍.影响小麦产量因素对产量形成的贡献[J].江苏农业科学,2012,40(10):78-80.
[4] 王柳,熊伟,温小乐,等.温度降水等气候因子变化对中国玉米产量的影响[J].农业工程学报,2014,30(21):138-146.
[5] 王莹,樊鑫,李豪杰,等.基于深度网络的多形态人脸识别[J].计算机科学,2015,42(9):61-65.
[6] 夏春江,王培良,张媛.基于深度信念网络的木材含水率预测[J].杭州电子科技大学学报(自然科学版),2015(1): 31-35.
[7] 杨旭华,钟楠祎.基于深度信念网络的医院门诊量预测[J].计算机科学,2016,43(S2):26-30.
[8] 郑志蕴,李步源,李伦,等.基于云计算的受限玻尔兹曼机推荐算法研究[J].计算机科学,2013,40(12):259-263.