An approach towards the modeling of Pattern Similarity in Music using Statistical Measures

The paper proposes an approach that finds the similarity between two songs. This song pattern similarity can is established by knowing the note structures and their fundamental frequencies of each note of the two songs and then compute the Pearson's Correlation of Coefficient. If the value of Correlation of Coefficient is very near to 1, that indicates the two songs are very much similar, otherwise it indicates a certain percentage of similarity. To establish the work, consider some songs as the test data and find out the song pattern similarities among them. The main aim of this paper is that it can be used as the song pattern similarity matching of music in the field of pattern recognition. The primary aspect of this work is to find the efficiency of some statistical approaches like Pearson's Correlation Coefficient is used for the matching patterns of two different song structures of combination of frequencies of different note structures in the field of Computational Musicology.