Rafia Ghulam Ali2025-07-302025-07-302020https://escholar.umt.edu.pk/handle/123456789/4006Extantevidence shows that clustering analysis can improve portfolio selection and performance. This study aims to improve portfolio performance using K-means clustering technique in Pakistan Stock Exchange (PSX). For this purpose, using daily return for the period of January 2010 to December 2018, the study develops three stock clusters based on Firm Size, Return on Equity and Return on Assets. The findings reveal that the clustering technique out performs naïve portfolio strategies in traditional Markowitz mean-variance framework.en-USThesis