Method trees

Ingo Mierswa, Katharina Morik
2005 Proceedings of the 2005 workshops on Genetic and evolutionary computation - GECCO '05  
In this paper we introduce a framework for automatic feature extraction from very large series. The extracted features build a new representation which is better suitable for a given learning task. The development of appropriate feature extraction methods is a tedious effort, particularly because every new classification task requires tailoring the feature set anew. Therefore, the simple building blocks defined in our framework can be combined to complex feature extraction methods. We employ a
more » ... enetic programming approach guided by the performance of the learning classifier using the new representation. Our approach to evolve representations from series data requires a balance between the completeness of the methods on one side and the tractability of searching for appropriate methods on the other side. In this paper, some theoretical considerations illustrate the trade-off. After the feature extraction, a second process learns a classifier from the transformed data. The practical use of the methods is shown by two types of experiments in the domain of music data classification: classification of genres and classification according to user preferences.
doi:10.1145/1102256.1102322 dblp:conf/gecco/MierswaM05 fatcat:4qa5bf32kze6vl326rmjcp37vm