Improving E-commerce Text Detection: A Comparative Study of Hybrid Approaches

Abstract: Text recognition for E-Commerce boosts search engine performance, facilitates product discovery, enhances user experience, lets you create personalized recommendations, and simplifies inventory control. We’ve developed a novel Ensemble SVM Multinomial Naive Bayes Approach in our research project that is specifically designed to detect e-commerce text. Our dataset had four different classes: books, electronics, household, and clothing and accessories. It contained 50,425 numeric values. Our impressive training accuracy of 99.83% and validation accuracy of 98.35% were attained by applying this state-of- the-art model. The precision of e-commerce text detection has advanced significantly with this accomplishment. We truly believe that our detection technology is capable of what it does. In our opinion, it offers a sound and useful approach to the analysis of text related to online commerce. In addition, we anticipate its integration serving as a cornerstone of future e-commerce sections, promising improved functionality and precision, which excites us about its future potential to refine the e-commerce scene. Index Terms: component, formatting, style, styling, insert.