Kategorien
September 2025 M D M D F S S 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Meta
Archiv der Kategorie: MIR Research
Runtime comparison of classifiers
The figure shows the runtime analysis of some popular classifiers for multi-objective feature selection identifying the piano occurrence in musical intervals. In my opinion it provides some very interesting details. X axis: hypervolume (minimizes the feature number and error). Y … Weiterlesen
Veröffentlicht unter MIR Research
Schreib einen Kommentar
IEEE SPS magazine article
An overview article which very well presents past and current SIGMA activities has been published in the July issue of IEEE Signal Processing Magazine: H. Blume, B. Bischl, M. Botteck, C. Igel, R. Martin, G. Rötter, G. Rudolph, W. Theimer, … Weiterlesen
Veröffentlicht unter Publications
Schreib einen Kommentar
Presentation of tempo recognition thesis at AES conference
The paper which summarizes the results of the diploma thesis „Tempo recognition based on audio signal for slow music pieces“ is accepted for the proceedings of the AES conference and will be presented as a poster. (Thorsten Deinert, Igor Vatolkin, … Weiterlesen
Veröffentlicht unter Publications
Schreib einen Kommentar
GECCO paper accepted
A paper is accepted for the GECCO conference: I. Vatolkin, M. Preuß, G. Rudolph – Multi-Objective Feature Selection in Music Genre and Style Recognition Tasks
Veröffentlicht unter General, Publications
Schreib einen Kommentar
New EA in AMUSE
Several new evolutionary strategies developed for the publication mentioned in http://sig-ma.de/2010/87 are now available in AMUSE for optimization of feature selection. It is possible to adjust the mutation probability based on the correlation with the label and to manage the … Weiterlesen
Veröffentlicht unter AMUSE & MIR Software
Schreib einen Kommentar
Application of SMS-EMOA for recall/specificity optimization
After SMS-EMOA [1] became a part of the actual (currently unreleased) AMUSE version, I’ve run a couple of experiments optimizing both recall and specificity. Here is an example for the performance progress for Jazz music category (blue circles: solutions from evaluations 1-500; green: … Weiterlesen
Veröffentlicht unter MIR Research
Schreib einen Kommentar
RapidMiner R plugin
Data Mining framework RapidMiner has now connection to R! Many further statistical methods and even the application of any R scripts becomes possible. For further details see: http://rapid-i.com/content/view/219/1/lang,en/ In future we are going to integrate the new RapidMiner into AMUSE.
Veröffentlicht unter AMUSE & MIR Software, MIR Research
Schreib einen Kommentar
PPSN publication about feature selection by EA
The joint publication of the Faculty of Computational Statistics and Chair of the Algorithm Engineering (TU Dortmund) [1] was recently presented as poster at the PPSN conference in Krakow. Evolutionary strategy driven by asymmetric mutation and several hybridization methods were developed and compared for the feature selection task for … Weiterlesen
Veröffentlicht unter Publications
Schreib einen Kommentar
New book about MIR
A new interesting book has been published recently at Springer: Ras, Zbigniew W.; Wieczorkowska, Alicja (Eds.): Advances in Music Information Retrieval (http://www.springer.com/engineering/mathematical/book/978-3-642-11673-5). It will be soon available in our library at Chair of Algorithm Engineering.
Veröffentlicht unter Publications
Schreib einen Kommentar
Bug fixed
The mapping of song partitions to songs was not done correctly during the n-fold cross-validation; now this bug has been fixed. Please checkout the new AMUSE version!
Veröffentlicht unter AMUSE & MIR Software
Schreib einen Kommentar