Scholarlink

The e-commerce industry faces a challenge in retaining its customer base due to cut-throat competition from multiple players and promotional offers. Today’s customers have plenty of options to choose from local as well as international players. This paper tries to explain the machine learning model to predict and prevent customer churn, thus improving sustainable practices for the e-commerce industry. Using the artificial neural network (ANN) model and other machine learning (ML) models, the current paper proposes to examine how customer churn prediction can be performed with the best accuracy. Preventing churn will lead to improved customer experience, less usage of financial resources, and better sustainable practices. The paper analyzed 11,260 e-commerce customer data published in credible sources. Multivariate analysis and ML models in Python have been used to find customer trends and prediction modelling. The study showed that demographic factors play an important role in churn. An active customer base, user-friendly mobile applications, e-payment or card usage, and high service scores ensure less customer churn. While predicting customer churn, 19 ML models were executed and the ANN model with multiple hidden layers (100,100,100,100) gave the perfect precision, recall, accuracy and F1 score. Managerial implications were discussed. The improvement and future research areas have also been mentioned.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Creating a Sustainable Future by Preventing Customer Churn in E-commerce—A Comparison of Machine Learning Models

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Swarnava Sengupta,
  • Amrin Samar Sultana,
  • Mrinmoy Bhattacharjee,
  • Abdul Hameed

摘要

The e-commerce industry faces a challenge in retaining its customer base due to cut-throat competition from multiple players and promotional offers. Today’s customers have plenty of options to choose from local as well as international players. This paper tries to explain the machine learning model to predict and prevent customer churn, thus improving sustainable practices for the e-commerce industry. Using the artificial neural network (ANN) model and other machine learning (ML) models, the current paper proposes to examine how customer churn prediction can be performed with the best accuracy. Preventing churn will lead to improved customer experience, less usage of financial resources, and better sustainable practices. The paper analyzed 11,260 e-commerce customer data published in credible sources. Multivariate analysis and ML models in Python have been used to find customer trends and prediction modelling. The study showed that demographic factors play an important role in churn. An active customer base, user-friendly mobile applications, e-payment or card usage, and high service scores ensure less customer churn. While predicting customer churn, 19 ML models were executed and the ANN model with multiple hidden layers (100,100,100,100) gave the perfect precision, recall, accuracy and F1 score. Managerial implications were discussed. The improvement and future research areas have also been mentioned.