Detecting Counterfeit Fashion Products Online using AI and Web Scraping
Main Article Content
Abstract
An important danger to brand integrity, income, and customer confidence is the widespread presence of fake luxury fashion items in online marketplaces. This study suggests an artificial intelligence (AI) framework that combines picture recognition, natural language processing (NLP), and web scraping methods to detect and flag fake listings on e-commerce platforms. A machine learning (ML) algorithm is used in the present study to identify counterfeit customer evaluations in a dataset. To determine the legitimacy of these reviews and then evaluate their validity, the program makes use of predictive modelling methods. The method employs convolutional neural networks (CNNs) to identify visual differences between genuine and fake products and uses natural language processing (NLP) to examine product evaluations and descriptions for questionable trends after automatically gathering product data and photos from many websites. In the current day, this procedure is especially crucial in domains like business intelligence. In this work, we suggest an AI system that works as a personal assistant for a designer of fashion products. The architecture of the system and all of its parts are shown, with special attention paid to the subsystems for data gathering and data clustering. Our use case scenario involves retrieving datasets of clothing items from two distinct sources and using Natural Language Processes to convert them into a certain format. Comparative results are shown after the two datasets are grouped independently using various mixed-type clustering algorithms, demonstrating the value of the clustering process in the issue of recommending apparel products.