A Commodity Search System for Online Shopping Using Web Mining

Online shopping and browsing are made convenient for customers by the rapidly expanding global e-commerce industry. However, different websites frequently charge different amounts for the same things, which leads to wasteful spending. Online shoppers must overcome the difficulty of devoting a significant amount of time and energy to finding the greatest offers and discounts, even though it is convenient and accessible. Although it can take time, manually filtering and comparing data can still produce ambiguous findings. Using web mining techniques, this work proposes a commodity search system for online shopping. Product data is obtained from well-known online stores like Amazon and Flipkart by using web scraping with Cheerio. Modern recommendation models like TF-IDF, Word2Vec, BERT, and Skip-Gram are used to improve user experience and support decision-making. These models' accuracy is determined to be 82.42%, 94.47%, 96.72%, and 97.35%, respectively, by thorough study. By giving consumers customized recommendations, this approach simplifies the online buying process and eventually helps consumers make wise purchases