Latent features based prediction on new users' tastes

Ming Chu Chen, Hung-Yu Kao

Research output: Contribution to journalConference articlepeer-review


Recommendation systems have become popular in recent years. A key challenge in such systems is how to effectively characterize new users' tastes - an issue that is generally known as the cold-start problem. New users judge the system by the ability to immediately provide them with what they consider interesting. A general method for solving the cold-start problem is to elicit information about new users by having them provide answers to interview questions. In this paper, we present Matrix Factorization K-Means (MFK), a novel method to solve the problem of interview question construction. MFK first learns the latent features of the user and the item through observed rating data and then determines the best interview questions based on the clusters of latent features. We can determine similar groups of users after obtaining the responses to the interview questions. Such recommendation systems can indicate new users' tastes according to their responses to the interview questions. In our experiments, we evaluate our methods using a public dataset for recommendations. The results show that our method leads to better performance than other baselines.

Original languageEnglish
Pages (from-to)176-187
Number of pages12
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8443 LNAI
Issue numberPART 1
Publication statusPublished - 2014 Jan 1
Event18th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2014 - Tainan, Taiwan
Duration: 2014 May 132014 May 16

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'Latent features based prediction on new users' tastes'. Together they form a unique fingerprint.

Cite this