上海水蜜桃病虫害模糊时间序列预测模型建立与验证  

Establishment and verification of fuzzy time series prediction model of honey peach diseases and pests in Shanghai

作　　者：王兴旺[1] 郑汉垣[2] 卢勇 Wang Xingwang;Zheng Hanyuan;Lu Yong(Shanghai Vocational College of Agriculture and Forestry,Shanghai 201699,China;School of Computer Engineering and Science,Shanghai University,Shanghai 200444,China;Shanghai Nanhui Peach Planting Base,Shanghai 201300,China)

机构地区：[1]上海农林职业技术学院,上海201699 [2]上海大学计算机工程与科学学院,上海200444 [3]上海南汇水蜜桃种植基地,上海201300

出　　处：《中国植保导刊》2022年第4期38-43,共6页China Plant Protection

基　　金：国家自然科学基金面上项目(61873156);国家自然科学基金重大研究计划重点项目(91630206)。

摘　　要：为提高上海水蜜桃病虫害的预测精度,笔者深入研究了模糊时间序列预测模型(FTSPM),在模糊C均值聚类算法(FCM)的基础上提出了低噪声模糊C均值聚类算法(LNFCM),从模糊聚类算法调用、历史数据论域划分、模糊关系建立等方面进行了改进和创新,建立了基于LNFCM的自调整模糊时间序列预测模型(LNFCM-SA-FTSPM),并应用上海南汇水蜜桃流胶病发生数据进行了验证。结果表明LNFCM-SA-FTSPM的整体预测性能、模型稳定性、预测准确率和精度都要优于常用的FTSPM模型,能够更好地指导上海水蜜桃病虫害防治工作。To improve the prediction accuracy of peach diseases and pests in Shanghai, we studied the fuzzy time se ries prediction model(FTSPM) in depth and proposed a low-noise fuzzy c-means clustering algorithm(LNFCM) based on the fuzzy c-means clustering algorithm(FCM). We improved the conventional model in the aspects of fuzzy clustering algorithm calling, domain division of historical data, the establishment of fuzzy relationships, and so on. A self-adjusting fuzzy time series prediction model based on LNFCM(LNFCM-SA-FTSPM) was established and verified by the occurrence data of honey peach gummosis diseases in Nanhui, Shanghai. The results showed that the overall prediction performance, model stability, prediction accuracy and precision of LNFCM-SA-FTSPM were better than those of the commonly used FTSPM model, it was supposed to give better guidance to the prevention and control of peach diseases and pests in Shanghai.

关 键 词：上海水蜜桃 病虫害 预测模型 聚类 

分 类 号：S431[农业科学—农业昆虫与害虫防治] S436.621[农业科学—植物保护]