A fault tolerant peer-to-peer distributed EM algorithm

Author

Safarinejadian, Behrooz ; Menhaj, Mohammad B. ; Karrari, Mehdi

Author_Institution

Electr. Eng. Dept., Amirkabir Univ. of Technol., Tehran

fYear

2009

fDate

March 30 2009-April 2 2009

Firstpage

46

Lastpage

52

Abstract

In this paper, a distributed Expectation Maximization (EM) algorithm is proposed for estimating parameters of a Gaussian mixture model in a peer-to-peer network. This algorithm is used for density estimation and clustering of data distributed over nodes of a network. Scalability and fault tolerance are two important advantages of this method. In the E-step of this algorithm, each node calculates local sufficient statistics using its local observations. A peer-to-peer algorithm is then used to diffuse local sufficient statistics to neighboring nodes and estimate global sufficient statistics in each node. In the M-step, each node updates parameters of the Gaussian mixture model using the estimated global sufficient statistics. The proposed method is then used for environmental monitoring and also distributed target classification. Simulation results approve promising performance of this algorithm.

Keywords

Gaussian processes; fault tolerance; optimisation; pattern classification; pattern clustering; peer-to-peer computing; statistical analysis; Gaussian mixture model; density estimation; distributed expectation maximization algorithm; distributed target classification; fault tolerant peer-to-peer computing; local sufficient statistic; pattern clustering; Association rules; Clustering algorithms; Data mining; Distributed databases; Fault tolerance; Gaussian distribution; Gaussian processes; Parameter estimation; Peer to peer computing; Statistical distributions;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Intelligence and Data Mining, 2009. CIDM '09. IEEE Symposium on

Conference_Location

Nashville, TN

Print_ISBN

978-1-4244-2765-9

Type

conf

DOI

10.1109/CIDM.2009.4938628

Filename

4938628