基于R语言的虫情危害预测算法设计

摘要
水稻是我国南方地区的主要粮食作物，在其种植过程中，水稻螟虫连年发生，极大威胁了我国水稻生产。预测害虫未来的发生动态，可以使治虫工作得以有目的、有计划、有重点的进行。只有对害虫发生危害的预测预报做到及时、准确，才能正确的拟定综合防治计划，及时采取必要的措施，经济有效的压低害虫的发生数量，以保证水稻的产量。
以根据预测害虫发生动态的结果进行有效防控为目的，论述一个基于R语言的害虫对水稻危害的预测算法，算法主要从影响害虫生存繁殖的月平均气温、月最低气温、月平均日照时间及月降雨量四个重要气候因素进行设计并实现，最终使得水稻害虫得到彻底防治。
实现了水稻害虫对水稻危害发生的动态的预测，使后期对水稻害虫的防治工作有较为准确的数据依据，从而达到对害虫有效防治目的。
关键词：R语言；预测；水稻害虫
Abstract
Rice is the main food crop in the south of China. In the process of rice cultivation, rice insects occur year after year, which greatly threatens the rice production in China. Predicting the future occurrence and dynamics of pests can make the pest control work carried out with a purpose, a plan and a focus. Only by timely and accurate forecast of pest damage can comprehensive control plan be correctly formulated and necessary measures be taken timely to economically and effectively reduce the number of pest occurrence so as to ensure the rice yield.
This paper mainly based on the results of the dynamic forecast pests for effective prevention and control, for the purpose of this paper discusses a R language based on rice pest hazard prediction, this algorithm mainly influence from pests survive the average temperature and lowest air temperature, sunshine time and rainfall four important climate factor to carry on the design and implementation eventually making rice pest prevention and control of radically, so that the rice yield stability.
Finally, a prediction algorithm was developed to predict the dynamics of rice pest damage to rice, so as to provide more accurate data basis for later rice pest control, so as to improve or guarantee rice yield.
Key words: R language; To predict; Rice pests
目录
1绪论 1
1.1 选题背景 1
1.2 主要内容及研究意义 1
1.3 论文组织结构 2
2水稻害虫特征概括及防治措施 3
2.1水稻害虫的分类、特征及其生存环境 3
2.2 水稻害虫的防治措施 4
3设计算法的方法简介 6
3.1线性回归模型 6
3.2 判别分析 7
3.3 BP神经网络 7
4数据处理与分析 9
4.1数据预处理 10
4.2线性回归模型分析 11
4.2.1 一元线性回归分析 11
4.2.2 多元线性回归分析 14
4.3判别分析 17
4.3.1 模型建立 17
4.3.2 模型优化 19
4.4 BP神经网络分析 20
4.4.1 模型建立 20
4.4.2 模型优化 22
4.4.3 建模重建 25
5预测模型结果及分析 27
5.1线性回归模型预测结果 27
5.2 判别分析预测结果 27
5.3 BP神经网络预测结果 28
5.4三种模型预测结果对比分析 29
6总结 30
致谢 31