Model for Predicting Success Factors in Agriculture in Kanchanaburi with Decision Trees

Abstract

This research aims to create a model predicting success from agriculture factors of Kanchanaburi province. Using a decision tree tool includes Random Forest Random Tree J48 Logistic Model Tree algorithm for the most effective algorithm for model and using factor selection by Feature method. Selection by Feature Elimination.
Results showed that the highest algorithm performance — J48, and when the Feature Selection — eliminate the factor that led to the highest predictive prototype performance at 5 factors, two patterns at the same accuracy of 70.3226.