Innovative Hybridisation of Genetic Algorithms and Neural Networks in Detecting Marker Genes for Leukaemia Cancer

Authors: Tong, D.L., Phalp, K.T., Schierz, A.C. and Mintram, R.

Conference: 4th IAPR International Conference in Pattern Recognition for Bioinformatics

Dates: 7-9 September 2009

Publisher: PRIB

Abstract:

Methods for extracting marker genes that trigger the growth of cancerous cells from a high level of complexity microarrays are of much interest from the computing community. Through the identified genes, the pathology of cancerous cells can be revealed and early precaution can be taken to prevent further proliferation of cancerous cells. In this paper, we propose an innovative hybridised gene identification framework based on genetic algorithms and neural networks to identify marker genes for leukaemia disease. Our approach confirms that high classification accuracy does not ensure the optimal set of genes have been identified and our model delivers a more promising set of genes even with a lower classification accuracy

http://eprints.bournemouth.ac.uk/13336/

Source: Manual

Preferred by: Keith Phalp

Innovative Hybridisation of Genetic Algorithms and Neural Networks in Detecting Marker Genes for Leukaemia Cancer

Authors: Tong, D.L., Phalp, K.T., Schierz, A.C. and Mintram, R.

Conference: 4th IAPR International Conference in Pattern Recognition for Bioinformatics

Publisher: PRIB

Abstract:

Methods for extracting marker genes that trigger the growth of cancerous cells from a high level of complexity microarrays are of much interest from the computing community. Through the identified genes, the pathology of cancerous cells can be revealed and early precaution can be taken to prevent further proliferation of cancerous cells. In this paper, we propose an innovative hybridised gene identification framework based on genetic algorithms and neural networks to identify marker genes for leukaemia disease. Our approach confirms that high classification accuracy does not ensure the optimal set of genes have been identified and our model delivers a more promising set of genes even with a lower classification accuracy

http://eprints.bournemouth.ac.uk/13336/

Source: BURO EPrints