Ecommerce websites today have become massive marketplaces of several hundred million products. Rakuten Ichiba alone hosts more than 300 million products. Such large marketplaces present numerous challenges in organizing all the product information so that customers are able to find products in an efficient manner. In this talk, I will describe how large-scale machine learning is helping us to organize and search this vast collection of information.
Specifically, I will talk about two of our projects in product categorization and search relevance. Product categorization refers to the task of organizing product information into appropriate genres of a taxonomy, and a good product categorization helps customers navigate a website correctly and find products efficiently. Additionally, I will briefly talk about our work on search relevance that uses feedback based on customer clicks to improve search algorithm in an ongoing manner.
Slide Deck
https://www.slideshare.net/secret/1PZrqUzogVCwqn