农作物害虫“0,1”型预报的因子筛选方法

• 作 者: 朱伯承
• 机 构: 中国科学院研究生院
• 出 处: 昆虫知识
• 发表日期: 1980年06期
• 关键词: 农作物害虫 因子 合率 拟合 预报量 预测预报 环境因素 预报模式 预报能力 害虫

影响农作物害虫发生发展的环境因素很多(马世骏,1978)。如何选出对害虫发生有密切联系的主要因子是作好害虫预测预报的重要一环。以前研究过的符号法(朱伯承,1978)虽有简单易作的优点,但选择的结果是比较粗糙的。本文首先谈谈衡量因子优劣的标准——拟合率,接着提出筛选因子的一些基本原则,然后介绍“0,1”型预报的二种不同的因子筛选方法。

There are many environmental factors that affect the occurrence and development of crop pests (Ma Shijun, 1978). How to choose the main factors that are closely related to pest occurrence is an important part of pest prediction. The symbolic approach previously studied (Zhu Bocheng, 1978) has the advantage of being simple and easy to use, but the result of selection is rather crude. This article begins with a discussion of the standard-fit rate, which measures the pros and cons of factors, followed by some basic principles for screening factors, and then introduces two different screening methods for “0, 1” predictions.