Classification Predictive Maintenance Using XGboost with Genetic Algorithm

This study develops a condition classification system of compressor 103J and water pump systems which are key equipment in the ammonia production line, hence monitoring these two very important machines. In recent years, there many good intelligent machine learning algorithms XGboost is one them. However, it contains parameters performance model will be greatly affected by selection their combination technique. this paper, algorithm combined with genetic algorithm, called GA-XGboost, order to find best hyper classifiers makes classifier more efficient ensures proper functioning systems. Experiments show that GA-XGboost has improved accuracy dataset compared other like Support Vector Machine (SVM), Random Forest (RF) AdaBoost. Also experiments demonstrate improvement different crossover operators algorithm.