Scholarlink

<p>In the digital age, accurately predicting online shopping intentions, particularly purchasing intentions, is crucial for developing tailored marketing strategies and improving customer experiences. We developed a novel approach by integrating TabNet, a cutting-edge deep learning model optimized for tabular data, with machine learning algorithms to enhance predictive accuracy. Using a dataset from e-commerce user sessions, we applied preprocessing techniques to address class imbalance and feature selection, followed by a comprehensive evaluation of several models, including support vector machines (SVM), K-nearest neighbors (KNN), decision tree, random forest, XGBoost, AdaBoost, LightGBM, CatBoost, logistic regression, and multilayer perceptron (MLP). The proposed ensemble learning model, TabNet-KNN, outperformed existing models, achieving superior predictive accuracy. The results demonstrate the efficacy of the TabNet-based ensemble learning approach in improving the predictive performance of online shopping intentions.</p>

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

Predicting online shopping intentions using TabNet-based ensemble learning approach

  • 期刊论文

  • 发表日期：

    2026年4月15日

查看全文
  • Yang Zhen,
  • Lili Zhang,
  • Xiaoyan Zhu

摘要

In the digital age, accurately predicting online shopping intentions, particularly purchasing intentions, is crucial for developing tailored marketing strategies and improving customer experiences. We developed a novel approach by integrating TabNet, a cutting-edge deep learning model optimized for tabular data, with machine learning algorithms to enhance predictive accuracy. Using a dataset from e-commerce user sessions, we applied preprocessing techniques to address class imbalance and feature selection, followed by a comprehensive evaluation of several models, including support vector machines (SVM), K-nearest neighbors (KNN), decision tree, random forest, XGBoost, AdaBoost, LightGBM, CatBoost, logistic regression, and multilayer perceptron (MLP). The proposed ensemble learning model, TabNet-KNN, outperformed existing models, achieving superior predictive accuracy. The results demonstrate the efficacy of the TabNet-based ensemble learning approach in improving the predictive performance of online shopping intentions.