Semantic Organization of Collaborative Users' Annotations Applied in Recommender Systems

• Grant number: 16/20280-6
• Support Opportunities: Regular Research Grants
• Start date: February 01, 2017
• End date: January 31, 2019
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Marcelo Garcia Manzato
• Grantee: Marcelo Garcia Manzato
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

Recommender systems aim to select and present relevant content according to users' preferences, reducing the information overload problem. Among the available techniques, the most known are collaborative and content-based filtering. In addition, there is a trend nowadays to use annotations collaboratively provided by users, such as tags, reviews, comments and interactions, in order to deal with common problems in recommendation, such as overspecialization, cold start and limited content analysis. However, these annotations may contain noise, irony and ambiguity, and may be represented in a unstructured and nonstandardized way. Moreover, there is a lack of a semantic organization of data so that they can be automatically infered to obtain the meaning of related concepts. Thus, this project aims to investigate how to use collaborative annotations produced by users to describe, in semantic terms, the entities involved in recommender systems. To reduce the problems related to using unstructured data, we plan to develop a method that uses different feature extraction, sentiment analysis and machine learning techniques to obtain a rich and standardized representation of items and users' preferences. (AU)

Scientific publications

(8)

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

DA COSTA, ARTHUR F.; MANZATO, MARCELO G.; CAMPELLO, RICARDO J. G. B.; ASSOC COMP MACHINERY. CoRec: A Co-Training Approach for Recommender Systems. 33RD ANNUAL ACM SYMPOSIUM ON APPLIED COMPUTING, v. N/A, p. 8-pg., 2018-01-01. (16/20280-6)

DA COSTA, ARTHUR; FRESSATO, EDUARDO; NETO, FERNANDO; MANZATO, MARCELO; CAMPELLO, RICARDO; ACM. Case Recommender: A Flexible and Extensible Python Framework for Recommender Systems. 12TH ACM CONFERENCE ON RECOMMENDER SYSTEMS (RECSYS), v. N/A, p. 2-pg., 2018-01-01. (16/20280-6)

DA COSTA, ARTHUR F.; MANZATO, MARCELO G.; CAMPELLO, RICARDO J. G. B.. Boosting collaborative filtering with an ensemble of co-trained recommenders. EXPERT SYSTEMS WITH APPLICATIONS, v. 115, p. 427-441, JAN 2019. (16/20280-6)

FRESSATO, EDUARDO P.; DA COSTA, ARTHUR F.; MANZATO, MARCELO G.; IEEE. Similarity-based Matrix Factorization for Item Cold-Start in Recommender Systems. 2018 7TH BRAZILIAN CONFERENCE ON INTELLIGENT SYSTEMS (BRACIS), v. N/A, p. 6-pg., 2018-01-01. (16/20280-6)

DE AGUIAR NETO, FERNANDO S.; DA COSTA, ARTHUR F.; MANZATO, MARCELO G.; CAMPELLO, RICARDO J. G. B.. Pre-processing approaches for collaborative filtering based on hierarchical clustering. INFORMATION SCIENCES, v. 534, p. 172-191, SEP 2020. (16/04798-5, 13/07375-0, 16/20280-6)

D'ADDIO, RAFAEL M.; MARINHO, RONNIE S.; MANZATO, MARCELO G.. Combining different metadata views for better recommendation accuracy. INFORMATION SYSTEMS, v. 83, p. 1-12, JUL 2019. (16/20280-6)

DA COSTA, ARTHUR F.; D'ADDIO, RAFAEL M.; FRESSATO, EDUARDO P.; MANZATO, MARCELO G.; ASSOC COMP MACHINERY. A personalized clustering-based approach using open linked data for search space reduction in recommender systems. WEBMEDIA 2019: PROCEEDINGS OF THE 25TH BRAZILLIAN SYMPOSIUM ON MULTIMEDIA AND THE WEB, v. N/A, p. 8-pg., 2019-01-01. (16/20280-6)

D'ADDIO, RAFAEL M.; FRESSATO, EDUARDO P.; DA COSTA, ARTHUR F.; MANZATO, MARCELO G.; ASSOC COMP MACHINERY. Incorporating Semantic Item Representations to Soften the Cold Start Problem. WEBMEDIA'18: PROCEEDINGS OF THE 24TH BRAZILIAN SYMPOSIUM ON MULTIMEDIA AND THE WEB, v. N/A, p. 8-pg., 2018-01-01. (16/20280-6)