| 基于优化模糊神经网络权值的大豆虫害诊断 |
| |
| 引用本文： | 韩霄,杨宇姝,冯江,袁琦.基于优化模糊神经网络权值的大豆虫害诊断[J].农机化研究,2017(3):247-252. |
| |
| 作者姓名： | 韩霄 杨宇姝 冯江 袁琦 |
| |
| 作者单位： | 东北农业大学 电气与信息学院,哈尔滨,150030 |
| |
| 基金项目： | 黑龙江省自然科学基金项目(F201401) |
| |
| 摘 要： | 采用模糊神经网络应用于大豆虫害快速识别方法。首先选择我国北方地区具有代表性的食心虫等7种虫害作为输出,用数字化特征表示。依据危害方式、危害症状等8种性状对182个大豆虫害样品进行诊断,选择1 3 6个样本作为训练集,选择4 6个样本作为预测集。首先使用AHP层次分析法对权值进行调整;其次,依据最优参数分别建立BP神经网络和模糊神经网络模型。实验结果表明:选择模糊神经网络进行模型建立,共预测对4 4个样本,判定识别率高达9 5%,证明了模糊神经网络进行大豆虫害判别是可行的。
|
| 关 键 词： | 大豆虫害 AHP 层次分析法 模糊神经网络 BP 神经网络 |
Soybean Pest Diagnosis Based on Optimized Weights of Fuzzy Neural Network |
| |
| Han Xiao,Yang Yushu,Feng Jiang,Yuan Qi.Soybean Pest Diagnosis Based on Optimized Weights of Fuzzy Neural Network[J].Journal of Agricultural Mechanization Research,2017(3):247-252. |
| |
| Authors: | Han Xiao Yang Yushu Feng Jiang Yuan Qi |
| |
| Abstract: | To distinguish soybean pest rapidly by using fuzzy neural network. First of all,choosing the representative 7 kinds of insects as output such as soybean pod borer and so on. Using digital features to express. According to 8 kinds of characteristics, such as damage mode and damage symptom, 182 soybean pest samples were diagnosed, 136 samples were selected as training set, and 46 samples were used as prediction set. Firstly, using the AHP analytic hierarchy process to adjust the weights;Secondly, the BP neural network and Fuzzy neural network model are established respec-tively according to the optimal parameters. The experimental results show that the method of selecting the analytic hierar-chy process and Fuzzy neural network for modeling, a total of 44 samples, the identification rate is 95%, it is proved that this method is feasible. |
| |
| Keywords: | soybean pest AHP analytic hierarchy process fuzzy neural network BP neural network |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
