When you listen to music, have you ever noticed that some songs are similar to the other while some are quite different? Maybe this is a universal phenomenon, however, our work try to provide fascinating insights into the characteristics and development of music. We create the adjacency matrix of 5603 musical artists to demonstrate the relationship between them, then we draw a subnetwork to show the connection between the influencers and the followers of Classical music in the 1930s. After that, we try to quantify the influences and similarities between these artists, using the adjacency matrix and principal component analysis (PCA) method, and establishing two indicators: ‘influence’ and ‘dissimilarities’. To test the practicability of the previous indicators, we apply them to analyze the genres of music, including the a comparison of influences and similarities between and within genres, the developing process of genres and the relationship of different genres. Especially, we use the random forest method to work out what distinguishes a genre. We find that artists within genre are statitically more similar to artists between genre.

Ziang Li . Genre Analysis based on Principal Component Analysis and Random Forest. Academic Journal of Computing & Information Science (2021), Vol. 4, Issue 2: 24-28. https://doi.org/10.25236/AJCIS.2021.040205.

[1] Erdős, P.; Rényi, A. On the evolution of random graphs. Publ. Math. Inst. Hung. Acad. Sci. 1960, 5, 17–60.

[2] R.Albert and A.-L. Barabási, Rev. Mod. Phys. 74, 47 (2002), DOI: 10.1103/RevModPhys.74.47. Crossref, ISI, ADS, Google Scholar

[3] Tang, P.; Song, C.; Ding, W.; Ma, J.; Dong, J.; Huang, L. Research on the Node Importance of a Weighted Network Based on the K-Order Propagation Number Algorithm. Entropy 2020, 22, 364.

[4] Russell SJ, Norvig P (2009). Artificial Intelligence: A Modern Approach. Prentice Hall, 3rd edition.