Author: Purnima Das
Das, Purnima, 2024 Medical knowledge discovery with effective data mining techniques, Flinders University, College of Science and Engineering
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
To enhance the quality of artificial intelligence-based health and medical knowledge discoveries, improvements to both the accuracy and usability of health and medical data mining and disease diagnosis tools are necessary. Several data mining approaches, including association rule mining (ARM), clustering, and classification techniques, have been developed by numerous computer scientists to support healthcare researchers. However, many of these techniques are unable to extract cost-effective and automatic models from vast amounts of data without user-defined (and, therefore, subjective) thresholds. Hence, the accuracy and reliability of the results can be degraded as a consequence.
With the goal of improving health and medical risk prediction and increasing the accuracy of the results, this thesis focuses on enhancing ARM algorithms, particularly those used for mining health and medical data. This study proposes several algorithms for the improvement of ARM techniques, which have been shown to extract knowledge efficiently.
Additionally, several of the challenges behind preparing noisy data and extracting useful information from real-world data were addressed. The ARM technique has been developed and employed on several health and medical datasets. Utilising the promising results found from this study as evidence, this research aims to improve the efficiency and accuracy of health risk prediction through health and medical knowledge discovery.
Keywords: Data Mining, Artificial Neural Networks, Association Rule Mining, Particle Swarm Optimisation, Knowledge Discovery, Optimised Frequent Itemsets, Medical and Health Data, Multi-Level Disease Classification, Cancer Sub-types.
Subject: Engineering thesis
Thesis type: Doctor of Philosophy
Completed: 2024
School: College of Science and Engineering
Supervisor: John Roddick