Precipitation’s Level Prediction Based on Tree Augmented Naïve Bayes model

Abstract

At present, most of the precipitation’s level predictions use the laws of nature to build the mathematical model which contains one or more series level to carry out the numerical simulation, as thus to analyze the causes and consequences of the evolution. Bayesian model is one kind of the foregoing said. In the Bayesian classification model, Naive Bayes model is known for its stability and easy to operate, but the established precedent assumption tends to be inadmissible. So here the article proposed  a new precipitation’s level prediction model based on the tree Augmented Naïve Bayes(we called TAN model for short hereafter), which improve the original Naïve Bayes model defects and increase the association between the leaf nodes on the basis of the original model. And we use the Dongtai station, Jiangsu province meteorology data to test the new precipitation model. The results show that the new precipitation prediction model’s performance is superior to the traditional Naive Bayes model.

DOI : http://dx.doi.org/10.11591/telkomnika.v12i1.3997