Monitoring Complex Data Feeds Through Ensemble Testing

Abstract

Managing and monitoring multiple complex data feeds is a major challenge for data mining tasks in large corporations and scientific endeavors alike. The invention describes an effective method for flagging abnormalities in data feeds using an ensemble of statistical tests that may be used on complex data feeds. The tests in the ensemble are chosen such that the speed and ability to deliver real time decisions are not compromised.

1 . A method of monitoring multiple data feeds for abnormalities, the method comprising:
 selecting at least two statistical tests to form an ensemble of statistical tests;   applying the ensemble of statistical tests to the multiple data feeds; and   displaying the results of the at least two statistical tests to determine abnormalities.