Artificial Immune Systems: A novel data analysis technique inspired by the immune network theory

Timmis, Jon (2000) Artificial Immune Systems: A novel data analysis technique inspired by the immune network theory. Doctor of Philosophy (Ph.D.) thesis, Department of Computer Science. (The full text of this publication is not available from this repository)

The full text of this publication is not available from this repository. (Contact us about this Publication)

Abstract

This thesis presents a novel data analysis technique inspired by the natural immune system. Immunological metaphors were extracted, simplified and applied to create an effective data analysis technique. This thesis builds on foundations of previous work, extracts salient features of the immune system and creates a principled and effective data analysis technique. Throughout this thesis, a methodical and principled approach was adopted. Previous work, along with background immunology was extensively surveyed. Problems with previous research were identified and principles from immunology were extracted to create the initial AIS for data analysis. The AIS, through the process of cloning and mutation, built up a network of B cells that were a diverse representation of data being analysed. This network was visualised via a specially developed tool. This allows the user to interact with the network and use the system for exploratory data analysis. Experiments were performed on two different data sets, a simple simulated data set and the Fisher Iris data set. Good results were obtained by the AIS on both sets, with the AIS being able to identify clusters known to exist within them. Extensive investigation into the algorithm's behaviour was undertaken and the way in which algorithm parameters effected performance and results was also examined. Despite initial success from the original AIS, problems were identified with the algorithm and the second stage of research was undertaken. This resulted in the resource limited artificial immune system (RLAIS) which created a stable network of objects that did not deteriorate or loose patterns once discovered. Periods of stable network size were observed with perturbations of the network size. This thesis presents a successful application of immune system metaphors to create a novel data analysis technique. Furthermore, the RLAIS goes a long way toward making AIS a viable contender for effective data analysis and further research is identified for study.

Item Type: Thesis (Doctor of Philosophy (Ph.D.))
Uncontrolled keywords: artificial immune systems, data analysis, kohonen networks, SOFM, unsupervised machine learning
Subjects: Q Science > QA Mathematics (inc Computing science) > QA 76 Software, computer programming,
Divisions: Faculties > Science Technology and Medical Studies > School of Computing > Applied and Interdisciplinary Informatics Group
Depositing User: Mark Wheadon
Date Deposited: 12 Sep 2009 10:23
Last Modified: 12 Sep 2009 10:23
Resource URI: http://kar.kent.ac.uk/id/eprint/21989 (The current URI for this page, for reference purposes)
  • Depositors only (login required):