Frequent Pattern Discovery in Online Environment

J. Dvorak (Czech Republic)

Keywords

Machine Learning, Pattern Recognition, Online Algorithm, Node Expiration

Abstract

In this article we present the online pattern discovery algo rithm. The algorithm is capable to discover frequent pat terns in the records of online sensors and similar devices. Each record is processed only once and only the processed record is available at the time. The algorithm operates on the records of discrete valued attributes and is capable to operate on the infinite sets of records as well as the infi nite sets of their attributes values. Neither the number of records nor the number of attributes and their values needs to be known in advance.

Important Links:



Go Back