Uni-Logo
Databases and Information Systems
Sie sind hier: Startseite Teaching Lehrangebot Sommersemester 2017 Various Aspects of Recommender Systems
 

Various Aspects of Recommender Systems

Projects

The projects offered here are aligned with the lecture "Data Analysis and Query Languages" offered in the summer semester, which covers the main RSs techniques.

This page will be updated soon with more projects' descriptions.

1. Recommending complementary products

Source: Learning to identify complementary products from DBpedia. V. A. Arrascue Ayala, Trong-Nghia Cheng, Anas Alzoghbi, Georg Lausen.

Short description

Identifying the complementary relationship between products, like a cartridge to a printer, is a very useful technique to provide recommendations. These are typically purchased together or within a short time frame and thus online retailers benefit from it. Existing approaches rely heavily on transactions and therefore they suffer from: (1) the cold-start problem for new products; (2) the inability to produce good results for infrequently bought products; (3) the inability to explain why two products are complementary.

The goal of this project is to develop an algorithm to predict potential complementary products without relying on transactions. This will be integrated into RecRDF4J, a framework developed at out department.

Dataset: Amazon product data.

Pre-requisites: knowledge of RDF, recommender systems. Programming in Java, Python; use of Maven.

Recommended number of participants: 2 per group

Status: not assigned

2. Recommending new Wikipedia articles

Short description

This project focuses on the recommendation of new Wikipedia articles. These are articles that are not yet in the English Wikipedia, but which became relevant for Wikipedia and therefore should be suggested to Wikipedia editors, journalists who would like to write about new topics, etc. The candidates for new Wikipedia articles are extracted from a large-scale news stream and are classified for recommendation based on a supervised machine learning model (given features such as the frequency over the last hours). More information about the general approach can be found in On Emerging Entity Detection. An existing framework (in Java) for Wikipedia article recommendation will be extended in different regards.

Dataset: Wikipedia (already preprocessed), news articles (will be provided).

Pre-requisites: basic knowledge of supervised machine learning, good programming skills in Java, significant interest in information extraction from text (as subfield of Natural Language Processing).

Recommended number of participants: 2 per group

Status: not assigned

3. Recommending references for (scientific) texts

Short description

A recommender sytem should be developed which takes as input a text document and outputs a text document which is enriched by citations. This means that the output text contains citation markers (e.g., "[1]") and a reference section in which references are listed. The general idea is that the system suggests publications to the user and directly inserts those publications as references into the text.

Related systems that are already existing:
Wenyi Huang, Saurabh Kataria, Cornelia Caragea, Prasenjit Mitra, C. Lee Giles, and Lior Rokach. Recommending citations: translating papers into references. In 21st ACM International Conference on Information and Knowledge Management, CIKM '12, pages 1910--1914, Oct 2012.
Qi He, Jian Pei, Daniel Kifer, Prasenjit Mitra, and Lee Giles. Context-aware Citation Recommendation. In Proceedings of the 19th International Conference on World Wide Web, WWW '10, pages 421--430, 2010.
Wenyi Huang, Zhaohui Wu, Liang Chen, Prasenjit Mitra, and C. Lee Giles. A Neural Probabilistic Model for Context Based Citation Recommendation. In Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence, pages 2404--2410, 2015.
Wenyi Huang, Zhaohui Wu, Prasenjit Mitra, and C. Lee Giles. RefSeer: A citation recommendation system. In IEEE/ACM Joint Conference on Digital Libraries, JCDL 2014, pages 371--374, 2014.
Jie Tang and Jing Zhang. A Discriminative Approach to Topic-Based Citation Recommendation. In Proceedings of the 13th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD '09, pages 572--579, 2009.

Dataset: Publications and evaluation data set (will be provided).

Pre-requisites: good programming skills (in Java or similar), significant interest in information extraction from text (as subfield of Natural Language Processing).

Recommended number of participants:2-3 per group

Status: not assigned