Adaptive Estimation for the Distribution Model of Golden Apple Snail (Pomacea canaliculata (Lamarck)) Pests Using Kernel and Spline Smoothers with Goldenshluger-Lepski Method

Abstract

The accuracy of the golden apple snail pest distribution model estimation is very much needed by farmers in dealing with pest attacks, especially in the rainy season. This research aimed to obtain the best distribution model of golden apple snail pests with kernel estimators and spline smoothing through the Goldenshluger-Lepski adaptive bandwidth selection method with an estimation error rate below 10%. The parameters measured were population density 7-42 days after planting, Morisita index, and environmental correlation. The results showed that the population density of golden apple snail pests from four research locations differed significantly in both the juvenile phase (Pr>F = 0.00161), pre-adult (Pr>F = 0.000872), and adult (Pr>F = 0.019122). The highest density was found in Bandar Kedungmulyo District (9.23 individuals.m-2), while the lowest was found in Megaluh District (6.37 individuals.m-2). The population pattern is evenly distributed with a Morisita index of less than one and the highest index (Id = 0.469) was recorded in Megaluh District. The best population distribution model was obtained using the optimum h(7) kernel smoothing estimator, with the lowest Mean Square Error (0.001), and Mean Absolute Square Error (0.032) values in Megaluh District. Furthermore, the best distribution model was obtained using the natural cubic spline smoother with the lowest Mean Square Error (0.055), and Mean Absolute Square Error (0.020) values in Tembeleng District. In conclusion, the best golden apple snail pest distribution model was obtained using the adaptive kernel smoothing estimator of the Goldenshluger-Lepsky model approach, which produced the lowest estimation error rate compared to the spline smoother. This research contributes to developing the best distribution model for golden snail pests, which can strengthen the information technology database for monitoring, controlling, and utilizing the potential of golden snail pests.