E-Commerce Product Classification Using Fused Machine Learning

Abstract

-rapid commerce's growth has forced the development of automated product categorization systems to appropriately categories large volumes of products. This study presents a novel machine learning-based e-commerce product classification algorithm. Starting with dataset normalization utilizing better POS tagging with WordNet and word vector calculation using LSTM 128 units, the recommended approach contains several key steps. This step ensures the classification process is proper. An ensemble feature selection technique employing RF with LR and Elastic net with Bag of Words (BoW) improves accuracy. The feature selection process identifies the most important categorization features. RESNET150 and CNN-dense trains the normalized dataset. Accuracy assessment uses Decision Tree, Random Forest, and Support Vector Machine. These tests evaluate the recommended technique and choose the optimal machine learning algorithm for the categorization challenge. Combining machine learning methods yields the best categorization. Fusion combines data from several sources to get complete understanding of an item. To improve classification, we blend machine learning algorithm results. E-commerce companies may improve inventory management and customer experience by properly categorizing products, according to this research.