Year of Graduation
Implementing Content-Based Recommender System for Fashion Industry Using Deep Neural Networks
Applied Mathematics and Information Science
Recent advances in deep machine learning algorithms have shown superior effectiveness in a large variety of tasks versus conventional approaches. In this paper, we aim to research how these methods can be applied to solve the task specific to the fashion industry and as such we chose task of building content-based item-to-item recommender system. We start with an explanation of this choice. Then we describe key features of the available infrastructure and raw data that was provided by the company. After that, we report which methods we have researched and which of them we used in the implementation of the production-ready system.