SEO Optimization of Product Cards on the Marketplace

The final qualification work is devoted to the development of an effective and easy-to-use application that will allow users to easily and quickly select a product category for a marketplace based on its photo, as well as create an SEO description for it. The paper analyzes existing solutions, examines various architectures of neural networks, describes the tasks of classifying and creating captions to images, and develops solutions for implementing neural network models for SEO optimization of product cards, taking into account the specifics of the data. Special attention is paid to the metrics and data collection methods used. The result of the work is a developed service based on a neural network capable of classifying products on the marketplace based on their photos and generating descriptions corresponding to them. Further research in this area may include improvements in the quality of results, as well as expanding functionality by generating a description using keywords.