Abstract : Here we study the Marshall-Olkin formulation of bivariate Pareto distribution, which includes both location and scale parameters and find efficient estimation techniques of the parameters of corresponding distribution. We use Maximum likelihood Estimation through the EM algorithm for the parameter estimation. Pareto distribution is heavy-tail in nature. It plays an important role in the Extreme Value Theory. So these distributions can be very useful in modeling the data related to finance, insurance, climate and network-security etc. These distributions can be used to analyze data related to any bivariate-component systems, e.g. axial length of two eyes of a diabetic patient. A numerical simulation is performed to verify the performance of different proposed algorithms
Venue
SMS seminar room
Speaker
Biplab Paul
Affiliation
IIT G
Title
Some variations of EM algorithms for estimating parameters of singular Bivariate Marshall-Olkin Pareto distribution