Importance Of Data Mining In Astronomy

Data Mining in Astronomy
Omer bin Sohail,
Department of Computer Sciences , NUCES
Lahore Campus
[e-mail]
[e-mail]

Abstract—

INTRODUCTION

Astronomy is the study of celestial objects such as stars, galaxies, planets, moons, and nebulae and the physics, chemistry, and evolution of such objects. Over the years Astronomy has become an immensely data rich field and is growing in exponential rate. Over the last decade alone there has been an exponential rise in observed data; most of it is in digital form. This growth beckons for new powerful tools to analyze and summarize it, and Data mining is just the tool needed for it.
In this paper we will discuss Current state of Data mining in Astronomy relative to

This is why Data Mining has a somewhat mixed response from the researcher in this field. If used correctly, it can be a powerful tool, holding the potential to fully exploit the exponentially increasing amount of available data, promising great advances in Astronomy. However, if misused, it can be little more than the black-box application of complex computing algorithms that may give little insight, and provide questionable results.
Skepticism is not the only problem, now days there are Multi-Terabyte Sky Surveys and Archives which will soon reach Multi-Petabyte, Billions of Detected Sources, and Hundreds of Measured Attributes per Source. Below are the high lights of current trends of Observational Astronomy
• Large digital sky surveys are becoming the dominant source of data in astronomy: currently 100 TB in major archives, and growing

The algorithm is ran on data blindfolded.
The most common of unsupervised method in astronomy is k-mean algorithm. K-means clustering is an unsupervised method that divides data into clusters. The number of clusters must be initially specified, but since the algorithm converges rapidly, many starting points can be tested. The algorithm uses a distance criterion for cluster membership, such as the Euclidean distance, and a stopping criterion for iteration, for example, when the cluster membership ceases to change.
Another interesting algorithm that only recently been used, in astronomy is COBWEB hierarchical clustering algorithm.
Cobweb Algorithm
The COBWEB is an incremental conceptual hierarchical clustering algorithm that was developed by machine learning researcher Douglas H. Fisher in the 1980s for clustering objects in an object-attribute data set. The COBWEB algorithm yields a clustering dendrogram called classification tree that characterizes each cluster with a probabilistic description. Each node in a classification tree represents a class (concept) and is labeled by a probabilistic concept that summarizes the attribute-value distributions of objects classified under the