A Local Scalable Distributed Expectation Maximization Algorithm for Large Peer-to-Peer Networks

Abstract

This paper offers a local distributed algorithm for expectation maximization in large peer-to-peer environments. The algorithm can be used for a variety of well-known data mining tasks in a distributed environment such as clustering, anomaly detection, target tracking to name a few. This technology is crucial for many emerging peer-to-peer applications for bioinformatics, astronomy, social networking, sensor networks and web mining. Centralizing all or some of the data for building global models is impractical in such peer-to-peer environments because of the large number of data sources, the asynchronous nature of the peer-to-peer networks, and dynamic nature of the data/network. The distributed algorithm we have developed in this paper is provably-correct i.e. it converges to the same result compared to a similar centralized algorithm and can automatically adapt to changes to the data and the network. We show that the communication overhead of the algorithm is very low due to its local nature. This monitoring algorithm is then used as a feedback loop to sample data from the network and rebuild the model when it is outdated. We present thorough experimental results to verify our theoretical claims.

Multi-objective optimization based privacy preserving distributed data mining in Peer-to-Peer networks

“…Datta et al [7] presents an overview of this topic. Examples of scalable distributed P2P data mining algorithms include the association rule mining algorithm [26], k-Means clustering [8], top-l inner product identification [6], decision tree induction [3], expectation maximization [2] and more.…”

Learning Structure and Schemas from Heterogeneous Domains in Networked Systems: A Survey

“…The authors of [4] presented an algorithm for learning parameters of Gaussian mixture models (GMM) in large P2P environments that can be used for a variety of wellknown data mining tasks in distributed environments such as clustering, anomaly detection, target tracking, and density estimation, which are necessary for many emerging P2P applications in bio-informatics, web-mining and sensor networks.…”

Decentralized EM to Learn Gaussian Mixtures from Datasets Distributed by Features

“…While these works are often formulated as generic optimization problems, rather than designed for a specific learning task, they tend to be motivated by applications where data is distributed by examples (horizontally partitioned), as is made clear, for example, in Forero et al (2010). Closer to our work, multiple fully decentralized algorithms use EM to fit GMMs in horizontally partitioned setups, such as Nowak (2003); Kowalczyk and Vlassis (2005); Gu (2008); Forero et al (2008); Bhaduri and Srivastava (2009);Safarinejadian et al (2010); Weng et al (2011); Altilio et al (2019). Related density estimation tasks have also been considered (Hu et al, 2007;Hua and Li, 2015;Dedecius and Djurić, 2017).…”