Scalable Distributed Change Detection from Astronomy Data Streams using Local, Asynchronous Eigen Monitoring Algorithms

Abstract

This paper considers the problem of change detection using local distributed eigen monitoring algorithms for next generation of astronomy petascale data pipelines such as the Large Synoptic Survey Telescopes (LSST). This telescope will take repeat images of the night sky every 20 seconds, thereby generating 30 terabytes of calibrated imagery every night that will need to be coanalyzed with other astronomical data stored at different locations around the world. Change point detection and event classification in such data sets may provide useful insights to unique astronomical phenomenon displaying astrophysically significant variations: quasars, supernovae, variable stars, and potentially hazardous asteroids. However, performing such data mining tasks is a challenging problem for such high-throughput distributed data streams. In this paper we propose a highly scalable and distributed asynchronous algorithm for monitoring the principal components (PC) of such dynamic data streams. We demonstrate the algorithm on a large set of distributed astronomical data to accomplish well-known astronomy tasks such as measuring variations in the fundamental plane of galaxy parameters. The proposed algorithm is provably correct (i.e. converges to the correct PCs without centralizing any data) and can seamlessly handle changes to the data or the network. Real experiments performed on Sloan Digital Sky Survey (SDSS) catalogue data show the effectiveness of the algorithm.

Detecting concept change in dynamic data streams

“…Apart from the above mentioned applications the field of radio astronomy has opened up a vast opportunity for knowledge discovery on distant galaxies through the use of data mining. Radio Astronomy data represents an ultra high speed data stream, is highly noisy, has slow rate of concept change and has potentially high false positive rate (Das et al 2009;Borne 2007).…”

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

“…Local algorithms for P2P data mining include the majority voting and protocol developed by Wolff and Schuster [37]. Based on its variants, researchers have further proposed more complicated algorithms: facility location [22], outlier detection [7], meta-classification [25], eigen vector monitoring [8], multivariate regression [4], decision trees [5] and the generic local algorithms [36].…”