论文部分内容阅读
用“双重筛选逐步回归”法对广西钦州县松毛虫发生进行分析,因变量为有虫株率与虫口密度(第2、3个月),输入18个自变量,包括虫口原来密度及有虫株率、月份、诸气象要素、防治与否(防治与否和相对湿度、防治与否和有虫株率、防治与否和虫口密度)等因子的交互作用等。用4种筛选临界值进行计算,随F_x值的增大,剔除的自变量(因子)越多,保留的自变量越重要。结果表明,虫口原来状况最为重要,人为干扰(防治与否)以及防治与否和相对湿度的交互作用是影响虫数重要因子之一,气象因子中以日照时数最为重要。月份也是不可忽略的因子。对下两个月虫情进行初步预报时体会到,预报较大范围内虫口平均密度不可能很准确,意义也不大。用“双重筛选逐步回归”法可以预报害虫发生地区及该地发生量,这是更有意义的。
The “double screening step by step regression” method was used to analyze the occurrence of pine caterpillars in Qinzhou County, Guangxi Province. The dependent variable was the worm strain rate and the insect population density (the first and second months). Eighteen independent variables including the original population density Plant rate, month, weather factors, prevention and control (whether or not prevention and control and relative humidity, prevention and control and whether the rate of insect strains, prevention and control and insect population density) and other factors such as the interaction. The four screening thresholds were used for the calculation. As the F_x value increased, the more exclusion variables (factors), the more important the retained independent variables. The results showed that the most important status of the insect population was the most important one. The interaction between man-made disturbance (prevention and control) and control or not and relative humidity was one of the important factors affecting insect numbers. The sunshine duration was the most important factor in meteorological factors. The month is also a factor that can not be ignored. The first two months when the worm forecast for the first time realized that the average density of insects within a wide range of forecasts can not be accurate, meaning is not large. It is more meaningful to use the “double screening stepwise regression” method to predict the occurrence of pests in the area and the area.