TY - JOUR

T1 - A Weight-based Information Filtration Algorithm for Stock-correlation Networks

AU - Hosseini, Seyed Soheil

AU - Wormald, Nick

AU - Tian, Tianhai

PY - 2021/2/1

Y1 - 2021/2/1

N2 - Several algorithms have been proposed to filter information on a complete graph of correlations across stocks to build a stock-correlation network. Among them the planar maximally filtered graph (PMFG) algorithm uses 3n−6 edges to build a graph whose features include high frequency of small cliques and good clustering of stocks. We propose a new algorithm which we call proportional degree (PD) to filter information on the complete graph of similarities between stocks. Our results show that the PD algorithm produces a network showing better homogeneity with respect to cliques, as compared to economic sectoral classification than its PMFG counterpart regardless of the similarity measure used—the Pearson correlation coefficient or normalised mutual information (NMI). We also show that the partition of the PD network obtained through normalised spectral clustering (NSC) agrees better with the NSC of the complete graph than the corresponding one obtained from PMFG. Finally, we show that the clusters in the PD network are more robust with respect to the removal of random sets of edges than those in the PMFG network.

AB - Several algorithms have been proposed to filter information on a complete graph of correlations across stocks to build a stock-correlation network. Among them the planar maximally filtered graph (PMFG) algorithm uses 3n−6 edges to build a graph whose features include high frequency of small cliques and good clustering of stocks. We propose a new algorithm which we call proportional degree (PD) to filter information on the complete graph of similarities between stocks. Our results show that the PD algorithm produces a network showing better homogeneity with respect to cliques, as compared to economic sectoral classification than its PMFG counterpart regardless of the similarity measure used—the Pearson correlation coefficient or normalised mutual information (NMI). We also show that the partition of the PD network obtained through normalised spectral clustering (NSC) agrees better with the NSC of the complete graph than the corresponding one obtained from PMFG. Finally, we show that the clusters in the PD network are more robust with respect to the removal of random sets of edges than those in the PMFG network.

KW - Normalised mutual information

KW - PD network

KW - PMFG network

KW - Stock-correlation network

UR - http://www.scopus.com/inward/record.url?scp=85094323120&partnerID=8YFLogxK

U2 - 10.1016/j.physa.2020.125489

DO - 10.1016/j.physa.2020.125489

M3 - Article

AN - SCOPUS:85094323120

VL - 563

JO - Physica A: Statistical Mechanics and its Applications

JF - Physica A: Statistical Mechanics and its Applications

SN - 0378-4371

M1 - 125489

ER -