RMT Assessments of the Market Latent Information Embedded in the Stocks' Raw, Normalized, and Partial Correlations
Figure 2
Probability Distribution Function (PDF) of the eigenvalues of the similarity measures for the S&P dataset, for the raw correlation (a), affinity (b), and partial correlation (c). The distributions are presented using straight lines, and the distributions for the corresponding random matrices are presented using red circles. In the case of the raw correlation and affinity matrices, the first eigenvalue (principal eigenvalue (PE)) is significantly larger than the other eigenvalues (131 and 233, resp.). In the case of the partial correlation, we see that there are more eigenvalues that are larger than the largest theoretical eigenvalue (Table 2), and the principal eigenvalue is now much smaller (33).
