sig-ma.de

Two papers were accepted for upcoming conferences:

D. Jannach, L. Lerche, I. Kamehkhosh: Beyond „Hitting the Hits“ – Generating Coherent Music Playlist Continuations with the Right Tracks. Proceedings of the 9th ACM Conference on Recommender Systems (RecSys 2015), Vienna, 2015

I. Vatolkin, G. Rudolph, C. Weihs: Evaluation of Album Effect for Feature Selection in Music Genre Recognition. Proceedings of the 16th International Society for Music Information Retrieval Conference (ISMIR 2015), Malaga, 2015

This month two papers were accepted.

The first one will appear in post-proceedings of ECDA 2014 and compares audio features and playlist statistics for music classification:
I. Vatolkin, G. Bonnin, D. Jannach – Comparing Audio Features and Playlist Statistics for Music Classification

The second one is accepted for EvoMUSART 2015. The conference will take place on 8.4-10.4.2015 in Copenhagen, Denmark. The paper describes a method for the maximization of interpretability in music classification without a limitation to use high-level (semantic) features only:
I. Vatolkin, G. Rudolph, C. Weihs – Interpretability of Music Classification as a Criterion for Evolutionary Multi-Objective Feature Selection

The paper I. Vatolkin – Exploration of Two-Objective Scenarios on Supervised Evolutionary Feature Selection: a Survey and a Case Study (Application to Music Categorisation) is accepted for EMO 2015. The conference will take place on 29.3-1.4.2015 in Guimarães, Portugal.

The article G. Bonnin and D. Jannach – Automated Generation of Music Playlists: Survey and Experiments, ACM Computing Surveys, Vol. 47, No. 2, Article 26 is now available online.

The paper is a detailed survey on playlist generation. First, different generation algorithms and the different types of data they can use are classified, the common approaches for the evaluation of the quality of the generated playlists are reviewed. In a second part, the results of thorough experiments are presented in which several existing algorithms are compared using several evaluation approaches. The results suggest that the order of the tracks, artist information, and track popularity are three important criteria to take into account when generating playlists and that some existing evaluation approaches have their limitations.

The paper on the analysis of playlist characteristics (D. Jannach, I. Kamehkhosh, and G. Bonnin – Analyzing the Characteristics of Shared Playlists for Music Recommendation) is accepted for the 6th ACM RecSys Workshop on Recommender Systems and Social Web.