Associate Professor, Department of Civil Engineering, Lakehead University, Canada

Dr. Jian Deng is currently an associate professor in the Department of Civil Engineering at Lakehead University and an invited researcher in the Centre of Excellence for Sustainable Mining and Exploration of Canada. Dr. Deng is a licensed professional engineer (P.Eng.) in Canada and has over 20 years’ research, education, and industry experience. Dr. Deng is the author/co-author of more than 80 referred journal publications on the topics of structural reliability, extreme value analysis, and stability of structures in the areas of Civil and Geotechnical Engineering.

Speech Title: Extreme Quantile Estimation Using Entropy and Self-Determined Probability Weighted Moments

Abstract: The estimation of extreme quantiles corresponding to small probabilities of exceedance (POE) is commonly required in the flood frequency and risk analysis of engineering systems. Traditionally, the principle of maximum entropy or minimum cross-entropy is used for estimating the probability density function under specified moment constraints. In such analyses, consideration of higher order moments is crucial for accurate modelling of the distribution tail. Since the higher order moment estimates from a small sample of data tend to be highly biased and uncertain, the use of entropy quantile estimates in extreme value analysis is fairly limited. The present research is an attempt to overcome this problem via the use of self-determined probability weighted moments (SD-PWMs). This paper presents a general approach to the estimation of the quantile function of a non-negative random variable using the principle of entropy subject to constraints specified in terms of SD-PWM estimated from observed data. The performances of the SD-PWM based quantile function were compared with those based on conventional PWMs. The results show that substantial improvement can be achieved in efficiency and accuracy of quantile estimation by use of SD-PWMs over of conventional PWMs, especially sample data containing outliers.