Recent development of content-based recommender systems
Maurizio Ferrari Dacrema
DEIB Ph.D. student
DEIB - BIO 1 Room (building 21, first floor)
September 20th, 2018
6.00 pm
Research Line:
System architectures
DEIB Ph.D. student
DEIB - BIO 1 Room (building 21, first floor)
September 20th, 2018
6.00 pm
Research Line:
System architectures
Sommario
Recommender systems are a tool used to help the user find relevant results in big catalogues and their application is now widespread in all kinds of services, from e-commerce to news or multimedia. Many different algorithms and techniques are now available to exploit very heterogeneous information sources, among which is textual information already available online. Although publicly accessible data sources like Wikipedia are a rich information source, being able to effectively embed this information in a recommender system poses a series of challenges. One of the main issue is to determine the correct interpretation of a word, hence understanding its semantics, and discover connected concepts or entities from non structured text. The main focus of this seminar is to provide an overview of some techniques that can be applied to extract semantic information from the web in order to enrich the representation of an item and improve recommendation quality.
The seminar is based on SWAP research group's course at RecSys, 2017 summer school.
The seminar is based on SWAP research group's course at RecSys, 2017 summer school.