Analyzing the stock market based on the structure of kNN network

This paper systematically studies the structure of the financial kNN (k-nearest neighbor) network. First, we use the eigenvalues and eigenvectors of the financial correlation matrix to analyze the structure of the network. We find that the degree is related to the average correlation coefficient, and furthermore, it also has a relationship between the components of the eigenvector corresponding to the maximum eigenvalue. We apply existing research to confirm that the community structure of the kNN network can be used to cluster financial time series. Finally, empirical studies based on financial markets in three countries show that there is a high correlation between the community structure and dimensions. Therefore, this study shows that the structure of the financial kNN network is related to the properties of the correlation matrix, and it extracts a meaningful correlation structure.