Presently, significance of deep learning techniques starts to overlook the world of E-commerce with their endless customizable online shopping experience to the users. Though huge data is streaming constantly during online commerce, data imbalance problem is still unaddressed due to insufficient analytical algorithms to handle huge datasets for smooth outliers. This leads to high congestion in the network as well as the extraordinary cost problem during online commerce. The foremost objective of this work is to resolve the classification task of imbalance data and churn rate using hybrid geometric sampling and AdaBoost based deep learning classification approach that uses diverse solution to provide a balance among prediction, accuracy, precision, specificity, sensitivity, and usability of data in E-commerce. This proposed solution helps to reduce the data imbalance problem and prediction of churn as well as non-churn customers in E-commerce web links. The experimental analysis has been carried out for the proposed algorithm in accordance with conventional techniques to check the practicability of the algorithm in real time practice.

中文翻译：

---

基于混合几何采样和AdaBoost的深度学习方法可解决电子商务中的数据不平衡问题

当前，深度学习技术的重要性开始以其对用户无穷的可自定义在线购物体验来忽视电子商务的世界。尽管在线交易期间海量数据不断流传，但由于分析算法不足以处理海量异常数据的庞大数据集，数据不平衡问题仍未解决。这导致网络中的高度拥塞以及在线商务期间的非凡成本问题。这项工作的首要目标是使用混合几何采样和基于AdaBoost的深度学习分类方法来解决不平衡数据和客户流失率的分类任务，该方法使用多种解决方案在预测，准确性，准确性，特异性，敏感性和可用性之间取得平衡。电子商务中的数据。提出的解决方案有助于减少数据不平衡问题以及电子商务Web链接中客户流失和非客户流失的预测。根据常规技术对提出的算法进行了实验分析，以检验该算法在实时实践中的实用性。