Audio Classification Based on Closed Itemset Mining Algorithm

Abstract

Automatic audio classification is a major topic in the fields of pattern recognition and data mining. This paper describes a new rule-based classification method (classification rule extraction for audio data, cREAD) for multiclass audio data. Typically, rule-based classification requires much computation cost to find rules from large datasets because of combinatorial search problems. To achieve efficient and fast extraction of classification rules, we take advantage of a closed itemset mining algorithm that can exhaustively extract non-redundant and condensed patterns from a transaction database in a reasonable time. A notable feature of this method is that the search space of the classification rules can be dramatically reduced by searching for only closed itemsets that are constrained by “class label item.” In this paper, we demonstrate that our method is superior to other salient methods for accurately classifying a real audio dataset.