Research and analysis of music development based on k-means and PCA algorithm

The purpose of this article is to establish an algorithm model that can measure the influence of music, capture the evaluation index reflecting the influence of music, and extend the model to other fields such as politics, culture, and society. We have established a music influence-oriented network algorithm model based on influencers and followers, where each artist is a node, and each follower is a connection between artists. We define relative interaction strength indicators to help understand the entire network algorithm. In addition, we also used time, genre and other scales to further optimize the network algorithm. We first use the PCA algorithm to determine indicators that reflect music similarity, such as vitality, activity, popularity, overall loudness, etc. On this basis, an evaluation algorithm model based on cosine similarity is established to calculate music similarity values of different genres. In addition, we use the K-MEANS algorithm to normalize each feature index and sum its variance. Finally, we noticed that the similarity of artists within genres is higher than the similarity of artists between genres. We further analyzed the differences and influences within and between genres. Taking time as a distinction, a relative heat map of the interactive influence of genres is drawn. It is understood that certain genres will obviously have a certain influence over time. We summarize this model as an impact correlation analysis model. First, we choose a representative influencer. Then, based on the cosine similarity, we obtained the music similarity with the fans in batches, thus more intuitively concluded that the Internet celebrities did affect the respective artists. In addition, we combined the calculation of SPSS variance and selected different indicators to visualize the radar chart to understand the attractiveness differences of certain music features. We first select the musical characteristics with obvious changing trends, then locate the position of the changer in the music evolution process through the time distribution diagram of the corresponding work, and finally select the representative changer. We analyzed the change history of each indicator in the selected genre over time, and finally got the global directed network diagram. Based on the network algorithm model established in the previous question, we analyzed the background of the times and found that there is an interaction between music and the cultural environment. Finally, we also analyzed the advantages and disadvantages of the algorithm model, and discussed the application of the method in other fields.

Research on the Influence of Music

As for question 1, based on the directed relationship between influencers and followers, we building a network of musicians based on influential relationships. A Music Influence Evaluation Model (MIEM) was also established, and the model formula is shown in the text. We then select the top 200 artists in the “music influence” ranking to build a subnet. The larger the subnet node, the more lines are extended. Indicating that the node represents the musician’s influence is large and extensive. From the graph, we can see that Bob Dylan is influential, but the breadth of influence is not enough; Miles Davis influenced a wide range of music factions.As for question 2?We have developed a Music Similarity Evaluation Model (MSEM) to calculate the contribution parameters of fifteen different music metrics. Using fully connected neural networks combined with triple loss to solve the answer. According to the characteristics of Triple Loss, we can make the similar nodes in the space closer together and the dissimilar nodes further apart. After training, our neural network is able to distinguish artists very well. The results were obtained: artists within genres are far more similar than artists between genres, and a classification image of musicians from different genres was produced.As for question 3, a comparative plot of characteristics revealed that music genres also have their own particular musical characteristics. The comprehensive analysis concludes that the difference between genres is mainly reflected by the six features of valence, tempo, mode, key, acousticness, and instrumentalness, and this result is verified by k-means clustering. By plotting the percentage of influence as well as the change of musical characteristics, it was concluded that the influence of genres changes over time; some musical characteristics in genres also change over time. Finally, the similarity between each faction is calculated and plotted as a heat map, and the genres with high similarity must have interrelated relationships with each other.As for question 4, we have developed a Music Influence T-test Model (MITM). We hypothesized that “influencers” would not influence followers to create music, and a t-test using SPSS rejected the original hypothesis and concluded that “influencers” would influence followers to create music. Additionally, Contagious Evaluation Model(CEM) was also be created. We established the “contagious” index and calculated the Pearson correlation coefficients between “contagious” and 15 musical characteristics, and obtained the results: energy, loudness, and acousticness are more “contagious” than other characteristics. Results: energy, loudness and acousticness are more “contagious” than other features.As for question 5, a time series plot of the variation for each musical characteristic with year was plotted and the analysis yielded the following conclusion: There are characteristics that signify revolutions in musical evolution from these data. For example, the music after 1960s showed changes characterized by higher rhythmicity, faster tempo, and fewer spoken words. Based on these musical evolutionary changes, combined with the “musical influence” we calculated earlier, we select five musical change-makers: The Beatles, Bob Dylan, The Rolling Stones, Miles Davis and Jimi Hendrix.As for question 6, we combined musical influences to identify the most influential musicians in each genre in each era as dynamic influencers to represent the music of the genre in that period. Creating images of their musical characteristics over time and analyzing them in relation to the history of musical development led to the conclusion that an artist’s musical identity changes with technology, social development, and changes in genre representation?As for question 7, a Network Connectivity Evaluation Model(NCEM) was developed to measure which artists in the music network were heavily influenced by external factors during the time period. The first and middle of the 20th century were found to be highly connected online, and this period coincided with a period of social upheaval, with the Cold War, World War II, the Industrial Revolution, and the rapid development of the Internet having a great impact on music, from which many new musical styles were born.

Quantitative Analysis of Comprehensive Influence of Music Network Based on Logistic Regression and Bidirectional Clustering

This paper makes a quantitative analysis of the comprehensive influence of music networks. Firstly, 11 music features are selected from energy, popularity, and other aspects to build a comprehensive evaluation index of music influence, and the PageRank algorithm is used to quantify the music influence. Secondly, the multiobjective logistic regression is used to construct the music similarity measurement model and, combined with music influence and music similarity, to judge whether the influence of different musicians is the actual influence. Thirdly, the influence and similarity of the same music genre and different music genres are analyzed by using the two-way cluster analysis method. Finally, the lasso region is used for feature selection to obtain the change factors in the process of music evolution and analyze the dynamic changes in the process of music development. Therefore, this paper uses network science to build a dynamic network to analyze the similarity of music, the evolution process, and the impact of music on culture, which has certain research significance and practical value in the fields of music, history, social science, and practice.

Human and automated judgements of musical similarity in a global sample

While music information retrieval (MIR) has made substantial progress in automatic analysis of audio similarity for Western music, it remains unclear whether these algorithms can be meaningfully applied to cross-cultural analyses of more diverse samples. Here we collected perceptual ratings from 62 participants using a global sample of 30 traditional songs, and compared these ratings against both pre-existing expert annotations and state-of-the-art audio similarity algorithms. We found that different methods of perceptual ratings all produced similar, moderate levels of inter-rater reliability comparable to previous studies, but that agreement between human and automated methods was always low regardless of the specific methods used to calculate musical similarity. Our findings suggest that current MIR methods are unable to measure cross-cultural music similarity in perceptually meaningful ways. We propose future directions to enable meaningful automatic analysis of all the world’s music.