Forecasting model for the fourth generation of cotton bollworm in corn fields based on big data

Abstract

The monitoring and forecasting model was put forward based on support vector machine program. According to the data collection of the fourth generation occurrence degree of the corn bollworm in Shandong province from 1999 to 2013, the support vector regression (SVR) method was adopted to build the nonlinear correlation model between the occurrence degree of the fourth generation bollworm and the associated factors. The method and the model were tested and analyzed. The results showed that the SVR forecasting model for prediction was almost in accord with the actual insect occurrence situation. The mean absolute percentage error was 4.36%, and the actual and estimated value of the correlation coefficient was 0.960 6. It could provide effective and accurate guidance to the cotton bollworm control in corn fields.