Portfolio Selection Using Network Filtering Methods: A Graph Theoretic Approach

The study presents a case of portfolio selection, allocation, and optimization using network-based methods. The study data consisted of selected stocks from Indian equity market. The computational analysis performed explores the use of networks science methods such as MST, Threshold filtering, centrality analysis, and disparity filtering. The portfolio selected based on these methods are further constructed, optimized and the evaluated using series of metrics such as reliability of portfolio, realized risk, and degree of reduction in the effective portfolio dimension. This study thus gives a mix of strategies that can be effectively used for portfolio selections and for construction of systemic-risk tolerant portfolios.