Personalised modelling framework and systems for Gene Data analysis and Biomedical applications
The core focus of this research is at the development of novel information methods and systems based on personalised modelling for genomic data analysis and biomedical applications. It has presented a novel personalised modelling framework and system for analysing the data from different sources and discovering the knowledge through an evolving and adaptive way. The main idea of personalised modelling is based on the assumption that every data sample has its unique pattern only being represented by a certain number of similar samples with a small set of important features. The proposed personalised modelling system (PMS) is an integrated computational system that combines different information processing techniques, applied at different stages of the data analysis, e.g. feature selection, classification, discovering the interaction of genes, outcome prediction, personalised profiling and visualisation, etc. In summary, this research has presented the main following contributions: (1) This study has implemented the idea of personalised modelling framework (PMF) introduced by Kasabov; (2) This study has developed novel algorithms and methods for PMS; (3) I have addressed the issuess in personalised modeling for data analysis and proposed solutions; (4) I have analysed the proposed PMS on 6 types of cancer gene expression data; (5) This thesis has presented the case studies of 4 types of cancer gene expression data analysis; (6) This study proposed a method using a coevolutionary algorithm for personalised modeling to select features and optimise relevant parameters for data analysis; (7) I have applied the proposed PMS on a SNPs dataset for Crohn’s disease risk evaluation in a real world case study; (8) The thesis gives the future research directions for personalised modelling study.