Pendekatan Data-Centric AI dengan Confident Learning dan XGBoost untuk Prediksi Churn Nasabah

Abstract

Customer churn prediction is critical for businesses to reduce customer attrition and improve customer retention strategies. This research proposes a method to enhance the accuracy and effectiveness of customer churn prediction in the banking industry using confident learning and XGBoost algorithms. Confident Learning is employed to address the challenges of uncertain or mislabeled training data, while XGBoost is utilized to build an efficient prediction model. The dataset used consists of 1,000 customers from a Regional Bank. The predictive model was tested in six different scenarios, and by using the confident learning algorithm and XGBoost, a significant accuracy improvement was achieved, reaching 93.36% on the dataset used.