Research on Improved Xgboost Algorithm for Big Data Analysis of e-Commerce Customer Churn

With the increasing cost of acquiring new users for e-commerce enterprises, it has become an important task enterprises to actively carry out customer churn management. Therefore, based on distributed gradient enhancement library algorithm (XGBoost), this research proposes a big data analysis study churn. First, conducts evaluation segmentation and combines random forest (RF) build RF XGBoost prediction model Finally, verifies performance model. The results show that area under receiver operating characteristic curve (AUC) value, accuracy, recall rate, F1 value RF-XGBoost are significantly better than those RF, XGBoost, ID3 decision trees model; average output error is 0.42, relatively good, indicating proposed in smaller higher accuracy. It can make general assessment then provide support maintenance work enterprises. helpful analyze relevant factors affecting churn, Equationte targeted service programs, thus improving economic benefits