A prediction model for population occurrence of paddy stem borer (Scirpophaga incertulas), based on Back Propagation Artificial Neural Network and Principal Components Analysis

Paddy stem borer (Scirpophaga incertulas) is an important insect pest of rice. Damaged plants wither and the tassels die or become blanched and infertile. Severe infestation leads to greatly decreased grain production. Best control of damage requires accurate describing and forecasting of the population dynamics. This paper applies Principal Components Analysis (PCA) and Back Propagation (BP) Artificial Neural Network (ANN) methods to analyze historical data on population occurrence to find out a non-line relation between the pest occurrence and the meteorological factors and then, to build a prediction model. Population data were collected from 2000 to 2008 by light trapping at the Plant Protection Station of JianShui County, Yunnan and associated meteorological data were obtained from the JianShui County Meteorologic Observatory. The new model successfully forecasted paddy stem borer population occurrence in 2006, 2007 and 2008. Test results show that there exactly exists the non-line relation between the insect population occurrence and the meteorological factors. And the new prediction model, based on BP ANN and PCA, improved prediction accuracy compared with other methods.