The statistical applications of K-Nearest Neighbors, Back Propagation Neural Networks, Support Vector Machines, Decision Trees, Random Forests, and Gradient Boosting Algorithms in data classification

Abstract

The Internet is closely related to people’s daily life now, and various types of Internet industries have emerged, such as IT and finance. People can’t live without finance, so Internet finance has become a popular trend. Traditional banks that lack the convenience of Internet finance are in the predicament of losing customers.

This article uses six machine learning methods to predict and classify customer churn in a European bank. According to the results of preliminary descriptive statistics and visual analysis, we clean, process and transform the data. We use the processed data to establish K-Nearest Neighbor, Back Propagation Neural Network, Support Vector Machine, Decision Tree, Random Forest and eXtreme Gradient Boosting models. We select the optimal model by analyzing, evaluating and comparing the evaluation indicators.

By analyzing customer information and the optimal model, we improve bank’s competitiveness and influences by proposing strategies to the bank.