Traditional and folk melody classifier on culture style using Markov models and neural networks

Music plays a vital role in any culture despite whether it is primary or modern and it is a good indicator reflecting the nature of the culture where it has been produced. Music traditions were developed even in the pre-historic periods when people did not have a proper method of documentation and scores of music. Melodies of different culture styles exhibit immense differences. Analyzing those differences is essential in various fields particularly ethnomusicology which studies non-western music based on cultural context. This paper presents an attempt of culture based melody classification using pitch. With respect to that, a prototype has been developed in Java, which utilizes a Markov chains and Neural Networks. Experiments were conducted with several datasets which were chosen from the traditional music styles such as Indian and Japanese.