• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

Extraction of Visual Features for Recommendation of E-commerce Items

Student: Andreeva Elena

Supervisor: Dmitry I. Ignatov

Faculty: Faculty of Computer Science

Educational Programme: Data Science (Master)

Final Grade: 9

Year of Graduation: 2017

This work is focused on the task of extracting visual features from products images and their further use in context-based recommender system. We observed and analyzed methods of extraction of visual features currently used by major companies and online stores such as Amazon, Pinterest, Houzz.com and Flipkart. Deep Convolutional Neural Network is used as a primary method of extraction of visual features in this work. We proposed several variations of fine-tuning the net and also modifications of the net's structure, that were implemented in practice. We also proposed a method for forming personalized recommendations based on neural network. The experiments were conducted on an open data set «Amazon product data», containing information about the products and the behavior of users (the facts of purchase). The quality of the recommendations generated by the proposed system was evaluated in comparison with the «state-of-the-art» method of collaborative filtering. The best results shown by the proposed models surpassed the result of collaborative filtering on various metrics. Keywords: Convolutional Neural Networks, Recommender Systems, Extraction of Visual Features.

Full text (added May 30, 2017)

Student Theses at HSE must be completed in accordance with the University Rules and regulations specified by each educational programme.

Summaries of all theses must be published and made freely available on the HSE website.

The full text of a thesis can be published in open access on the HSE website only if the authoring student (copyright holder) agrees, or, if the thesis was written by a team of students, if all the co-authors (copyright holders) agree. After a thesis is published on the HSE website, it obtains the status of an online publication.

Student theses are objects of copyright and their use is subject to limitations in accordance with the Russian Federation’s law on intellectual property.

In the event that a thesis is quoted or otherwise used, reference to the author’s name and the source of quotation is required.

Search all student theses