UROP Openings

Have a UROP opening you would like to submit?

Please fill out the form.

Submit your UROP opening

Deep learning-based recommender system with natural language processing and deep learning using social media data


Term:

Spring-Summer

Department:

MAS: Media Arts and Sciences

Faculty Supervisor:

Kent Larson

Faculty email:

kll@mit.edu

Apply by:

February 2020

Contact:

Luis Alberto Alonso Pastor: alonsolp@media.mit.edu

Project Description

Many real-world data cover reviews and ratings on business, creating opportunities for marketing companies to better understand the demand and preferences of customers with complementary information. However, to effectively combine data with multi-modal nature and complex structure is challenging. Users post reviews on the businesses, revealing the characteristics of businesses they care. We aim to use users’ reviews as auxiliary information and their reviews to build personalized recommendations, which reveal what they care about most. We will develop a novel deep learning framework to utilize nodal feature information and graph structures effectively. Our framework will be able to extract the motif structure as well as removing uninformative features, which will enable us to understand what do users care about and how do businesses perform in different dimensions. The method will be evaluated on the scraped data, such as TripAdvisor, to demonstrate its effectiveness.

Pre-requisites

Skills you need: — Proficiency in Python (Required) — Machine learning (Required) — Experience in data scraping (Bonus) — Deep learning (Optional) — Social networks (Optional) Skills you will learn: — Graph-based deep learning — Behavioral data analysis