Any investor in stock markets around the world has a deep concern about the shortfalls of allocation wealth to any stock without accurate estimation of related risks. As we review the literature of risk management methods, one of the main pillars for the risk management framework in defining risk measurement approach using historical data is the estimation of the probability distribution function. In this paper, we propose a new measure by using kernel density estimation via diffusion as a nonparametric approach in probability distribution estimation to enhance the accuracy of estimation and consider some distribution characteristics, investor risk aversion and target return which will make it more accurate, compre-hensive and consistent with stock historical performance and investor concerns.
