EFFECTIVE EXPECTATION MAXIMIZATION ALGORITHM IMPLEMENTATION USING MULTICORE COMPUTER SYSTEMS
Journal Title: Informatyka Automatyka Pomiary w Gospodarce i Ochronie Środowiska - Year 2014, Vol 4, Issue 4
Abstract
A popular expectation maximization algorithm that is widely used in modern data processing systems to solve various problems including optimization and parameter estimation is considered in the paper. The task of the study was to enhance effectiveness of the algorithm execution in time.An enhancement of execution rate for the EM algorithm using multicore architecture of modern computer systems was carried out. Necessary modifications aimed at better parallelism were proposed for implementation of the EM algorithm. An efficiency of the software implementation was tested on the classic problem of Gaussian random variables mixture separation. It is shown that in the mixture separation problem EM algorithm performance degrades when the distance between mean values of distributions is less than three standard deviations, which is totally in the spirit of three sigma law.In such cases, it is very important to have an efficient EM algorithm implementation to be able to process such test cases in a reasonable time.
Authors and Affiliations
Alexei Kasitskij, Peter Bidyuk, Alexander Gozhyi
Keywords
Related Articles
- EP ID EP73248
- DOI 10.5604/20830157.1130179
- Views 104
- Downloads 0
INFLUENCE OF THE MAGNETIC FIELD ON FLOWING LIQUID OF SELECTED MAGNETIC PROPERTIES
This paper presents the results of the effect studies of the magnetic field on the paramagnetic, flowing liquid. In the area of direct magnetic field the eddy currents in flowing liquid are induced. The induced currents...
RESEARCH MADE IN INSTITUTE OF COMPUTER SCIENCE
In the article the research made in the Institute of Computer Science is presented. The most important areas are presented. Obtained results are described.
INVERSE PROBLEM SOLUTION USING NEURAL NETWORK
This paper was presented a practical method of solving inverse problems in tomography using neural networks. The methods to detect changes in the size and position of an object inside another object tomography. Computed...
METHODS OF PRODUCING APODIZED FIBER BRAGG GRATINGS AND EXAMPLES OF THEIR APPLICATIONS
The paper presents the principle of operation, the structure, applications and methods of producing uniform, chirped and blazed fiber Bragg gratings as well as long period gratings. Finally, several selected methods used...
OBJECT-RELATIONAL MAPPING TECHNOLOGIES IN PHP APPLICATIONS
This paper presents a comparison of the two most commonly used for PHP applications object-relational mapping technologies: Propeland Doctrine. The comparative analysis was made on the basis of the test application and...