Data Clustering with Artificial Innate Immune System Adding Probabilistic Behaviour Pathak Vishwambhar1*, Dhyani Praveen Dr.2,**, Mahanti Prabhat Dr.3,*** 1Assistant Professor, Birla Institute of Technology (Ranchi) Jaipur Campus, 27, MIA, Jaipur. 2Executive Director, Banasthali University, Jaipur Campus, C-62, Sarojini Marg, C-Scheme, Jaipur. 3Professor in Computer Science, University of New Brunswick, St. John, Canada. * e-mail id: pathakvishi@gmail.com
** e-mail id: dhyani_p@yahoo.com
*** e-mail id: pmahnti@gmail.com
Abstract adoption of Artificial Immune System (AIS) for clustering of data stream is appealing considering its features to address the underlying issues where as the natural phenomena are perceivably probabilistic. This paper implements innate immune systems metaphor for the task of clustering. The contribution of this work is to simulate probabilistic behaviour of the Dendritic Cells (DCs) during recognition of the antigens and danger signals using an infinite Gaussian mixture model. The developed algorithm is capable of automatic determination of number of spherical clusters and sub-clusters. The results of experiments have been found comparable with those of existing models. Top Keywords Artificial Immune Systems, Danger Theory, Clustering, Infinite Mixture Models, Gaussian Mixture. Top |