| |
 |
| Paper: |
Data Mining in Distributed Databases for Interacting Galaxies |
| Volume: |
347, Astronomical Data Analysis Software and Systems XIV |
| Page: |
350 |
| Authors: |
Borne, K. |
| Abstract: |
We present results from an exploratory data mining project to identify classification features of special classes of interacting galaxies (for example, infrared-luminous galaxies) within distributed astronomical databases. Using a variety of data mining techniques, interaction-specific features are learned, to distinguish this class of galaxies from a control sample of normal galaxies. Eventually, the corresponding rule-based feature model of that class of galaxies will be applied to the large multi-wavelength astronomical survey databases that are becoming available. This distributed data mining activity is a prototype science use case for the VO (Virtual Observatory). We specifically apply multi-archive multi-wavelength data to the problem. In a preliminary validation experiment, we recovered exactly the type of object that we hope to find automatically with our data mining tools: a distant hyper-luminous infrared galaxy (HyLIRG), the most luminous class of known galaxies. This particular galaxy was previously known, but we re-discovered it serendipitously. |
|
|
|
|
|
