Scholarlink

<p>E-commerce has rapidly transformed the retail landscape, leading to increased competition and the necessity for data-driven decision-making. Managing inventories, developing marketing plans, and raising customer happiness all depend on accurate sales forecasting. Standard predictive methodologies face complexities within standard e-commerce sales data, characterized by high dimensionality or variability. This is the motivation for this work. In this work, we propose an entirely new methodology for the prediction of e-commerce sales, using an Efficient Parallel Novel Narwhal Convolutional Attention Module Network (Eff-P2N-Con-AtMNet). In this work, the input dataset is from the E-commercial Customers dataset. The method is to pre-process the E-commerce Customers dataset using Shape-Aware Mesh Normal Filtering (SAMNF) in order to enhance the data’s quality, and to extract features using Multiple Discrete Orthonormal S-Transforms to properly characterize underlying processes. Selecting which features will be included in the final prediction model is another usage for Synergetic Fibroblast Optimization (SFO). Convolutional attention processes may be used by the proposed Eff-P2N-Con-AtMNet to optimize e-commerce sales forecast accuracy and efficiency. The proposed Eff-P2N-Con-AtMNet model produces exceptional results, including a minimal error rate of 0.1, computational cost and complexity of 0.1, accuracy of 99.91%, recall of 96.59%, precision of 98.87%, and F1-score of 96.84%. This unique framework offers meaningful guidance on improving sales forecasting, demand planning, and merchandising for E-commerce businesses by identifying peak sales periods, understanding customer habits, and designing promotional activities to enhance customer engagement and satisfaction.</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

E-commerce system for sale prediction using efficient parallel novel Narwhal convolutional attention module network

  • 期刊论文

  • 发表日期：

    2025年10月16日

查看全文
  • Rajendra Pujari,
  • K. Vinitha,
  • Nagendrakumar Turaga,
  • R. Giri Prasad

摘要

E-commerce has rapidly transformed the retail landscape, leading to increased competition and the necessity for data-driven decision-making. Managing inventories, developing marketing plans, and raising customer happiness all depend on accurate sales forecasting. Standard predictive methodologies face complexities within standard e-commerce sales data, characterized by high dimensionality or variability. This is the motivation for this work. In this work, we propose an entirely new methodology for the prediction of e-commerce sales, using an Efficient Parallel Novel Narwhal Convolutional Attention Module Network (Eff-P2N-Con-AtMNet). In this work, the input dataset is from the E-commercial Customers dataset. The method is to pre-process the E-commerce Customers dataset using Shape-Aware Mesh Normal Filtering (SAMNF) in order to enhance the data’s quality, and to extract features using Multiple Discrete Orthonormal S-Transforms to properly characterize underlying processes. Selecting which features will be included in the final prediction model is another usage for Synergetic Fibroblast Optimization (SFO). Convolutional attention processes may be used by the proposed Eff-P2N-Con-AtMNet to optimize e-commerce sales forecast accuracy and efficiency. The proposed Eff-P2N-Con-AtMNet model produces exceptional results, including a minimal error rate of 0.1, computational cost and complexity of 0.1, accuracy of 99.91%, recall of 96.59%, precision of 98.87%, and F1-score of 96.84%. This unique framework offers meaningful guidance on improving sales forecasting, demand planning, and merchandising for E-commerce businesses by identifying peak sales periods, understanding customer habits, and designing promotional activities to enhance customer engagement and satisfaction.