Analysis of Gradient Boosting, Adaboost, Catboost Algorithms in Water Quality Classification

Abstract

This research aims to find the highest accuracy of the three classification algorithms. The highest accuracy algorithm will be used as a reference in this water quality classification. and test the performance of the third model. The method used in this analysis to overcome missing data is the median method. Then to handle unbalanced data, the SMOTE method is used. In this study, we compared the accuracy and performance of Gradient Boosting, Adaboost, and Catboost. The results found that the Catboost algorithm has the highest accuracy and performance of 68%, followed by Gradient Boosting at 60% and Adaboost at 58%. Then the performance of the AUC Catboost value is 0.678, Gradient Boosting is 0.595, and Adaboost is 0.584. But the results of accuracy and performance are still lacking.