Hierarchical Clustering of Music Database Based on HMM and Markov Chain for Search Efficiency

• Published in: Speech, Sound and Music Processing: Embracing Research in India pp. 98 - 103
• Main Authors: Ross, Joe Cheri, Samuel, John
• Format: Book Chapter
• Language: English
• Published: Berlin, Heidelberg Springer Berlin Heidelberg
• Series: Lecture Notes in Computer Science
• Subjects: content based information retrieval; Hidden Markov Model; information retrieval; Markov Chain; Music; music search; Scale based search
• DOI: 10.1007/978-3-642-31980-8_9
• ISBN: 9783642319792; 3642319793
• ISSN: 0302-9743

Music search unlike the regular text search works on huge databases and traditional pattern matching approaches are not feasible. The efficiency of a music search engine solely depends on the data categorization scheme employed. The proposed idea aims to reduce search complexity using tree based organization of music database and also considering scale, chord and note transition of the input query. Probabilistic modeling of chord transition by Hidden Markov model and notes transition through Markov chain improvise on clustering enormous music data, eventually resulting in search complexity reduction. The method inherently supports minor deviations in the input query which may prevent meeting user expectations despite the availability of data.