A Classifier Build Around Cellular Automata for Distributed Data Mining
Lianying Zhou, Mingxin Yang · 2008
A pattern classifying machine (PCM) ts-PCM for distributed data mining (DDM) was designed in this article. ts-PCM is based on a special class of linear cellular automata (CA), termed as mutiple attractor CA (MACA). Each MACA could be distributed in different sites as a base classifier. Characterization of a MACA based two stage by two linear operators of dependency string (DS) and dependency vector (DV) other than dependency matrix of the CA used to be, and employing genetic algorithm (GA) formulation. Its classification complexity has been declined from O(n3) to O(n). Plentiful experimental results have proved the potential of ts-PCM,and with the respect to excellent classification accuracy and low memory overhead established the availability of the classifier to manipulate the distributed data mining.