基于智能算法的月季鲜切花病虫害预测模型的探究  

Exploration on Chinese Rose Cut Flowers Diseases and Insect Pests Forecasting Model Based on Intelligent Algorithm

作　　者：钱晔[1,2] 孙吉红 彭琳 李文峰[2,4] 汪惜今 陆国泉 

机构地区：[1]云南农业大学基础与信息工程学院,云南昆明650201 [2]云南省高校农业信息技术重点实验室,云南昆明650201 [3]云南省科学技术院,云南昆明650000 [4]云南农业大学新农村发展研究院,云南昆明650201 [5]成都信息工程大学大气科学学院,四川成都610000

出　　处：《计算机技术与发展》2017年第12期157-160,165,共5页Computer Technology and Development

基　　金：国家自然科学基金资助项目(31260292);云南省自然科学基金(2012FD020);云南省教育科研基金项目(2015Y194);云南省教育社会科学基金(2012C086)

摘　　要：针对传统预测模型的主观性强、成本偏高、误差偏大等问题,设计并提出了基于自组织竞争神经网络算法的月季鲜切花病虫害预测模型。该模型能够有效预防病虫害危害,确保月季鲜切花的正常生长,从而确保云南省月季鲜切花的产量、质量和声誉。以最为典型的月季鲜切花白粉病为实例,通过问卷调查、头脑风暴法相结合的加权方法来确定影响因子的权重,并以60组影响因子的数据作为输入数据,建立了基于自组织竞争神经网络算法的病虫害预测模型。将所提出模型的预测结果与采用名义小组法所建立传统病虫害预测模型的预测结果进行对比分析。实验结果及其分析表明,基于自组织竞争神经网络算法的预测模型可有效地为月季鲜切花种植企业、农户、散户提供更加准确的信息,降低了种植的盲目性。In view of the problems of strong subjectivity,high cost and big error for traditional forecast model,based on self-organizing competitive ANN, the fresh cut flower plant diseases and insect pests forecasting model is designed. It can effectively prevent insect pests diseases from hazarding the normal growth of fresh cut flower and make sure the production,quality and reputation of fresh cut flowers in Yunnan Province. Taking most typical Chinese rose cut flowers powdery mildew for instance, the weights of impact factors are determined through questionnaire survey in combination with brainstorming, and then based on self-organizing competitive ANN, the forecast model is to be built with 60 sets of impact factor as input of it. Compared with the traditional plant diseases and insect pests forecasting model, the experimental results show that the proposed model can provide effectively more accurate information for the fresh cut flowers enterpri- ses, farmers, and retail investors to reduce their blindness of planting.

关 键 词：智能算法 月季鲜切花病虫害 自组织竞争神经网络 预测 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]