Compound Normal with Gamma Mixture Models - A Model Characteristic Perspective

Abstract

In this paper, we discuss convergence performance issues of Compound Normal with Gamma Mixture(CNGM) and Truncated Compound Normal with Gamma Mixture(TCNGM) models that we have proposed in our earlier work in comparison to the normal mixture model(NM). We show that these models are feasible for mixture density estimation. Based on examination of coefficient of kurtosis, these models are shown to be preserving model characteristics of the more mesokurtic normal distribution and model better the leptokurtic deviations of normal distribution.

Authors and Affiliations

Viziananda Row Sanapala, Sreenivasa Rao Kraleti, Srinivasa Rao Peri

Keywords

Related Articles

Computational Fluid Dynamics Analysis of Single Pass Shell & Tube Heat Exchanger with Different Orientation of Baffles and Without Baffles

In this specific project an endeavor is design for CFD evaluation of Single Pass Shell & Tube Heat Exchanger with Different Orientation of Baffles and Without Baffles with counter flow of fluid. In order to research the...

Customer Satisfaction in Construction Industry

The questions for the questionnaire survey were framed by considering the points given by both customers as well as the flat promoters.The questionnaire survey was conducted among two types of customers. The first type...

PSO Algorithm based Loss Minimization Approach for Optimal Placement and Sizing of Renewable Energy Source

In this research paper, PSO is joint with Newton Raphson method of power flow to optimize optimal location along with size of renewable energy source. A NR method is used to measure the losses and find the voltage at ea...

slugAnalysis of Various Image Steganography Techniques

Steganography is an art and science of hiding the information into some carrier files without changing the visual properties of carrier file. In digital world, the carrier file can be text file, audio/video file, image...

Design of Pseudorandom Pattern Generator for MIHST

The paper describes a system generator which produces Pseudo random sequence with user defined toggling range. The patterns were generated using PRPG; Detection of errors is done using Microprocessor Hardware Self Test...

Download PDF file
  • EP ID EP22954
  • DOI -
  • Views 242
  • Downloads 4

How To Cite

Viziananda Row Sanapala, Sreenivasa Rao Kraleti, Srinivasa Rao Peri (2016). Compound Normal with Gamma Mixture Models - A Model Characteristic Perspective. International Journal for Research in Applied Science and Engineering Technology (IJRASET), 4(12), -. https://europub.co.uk/articles/-A-22954