Zainuddin, Zarita and Kumar, Maragatham (2008) Radial Basis Function Neural Networks in Protein Sequence Classification. Malaysian Journal of Mathematical Sciences, 2 (2). pp. 195-204. ISSN 1823-8343
Applications of neural networks in bioinformatics have expanded tremendously in recent years due to the capabilities of neural networks to solve biological problems. Neural networks have been implemented in numerous biological fields. In this paper, standard radial basis function and modular radial basis function neural networks are used to classify protein sequences to multiple classes. n-gram method is used to transform protein features to real values. A learning strategy known as the selforganized selection of centers is presented. In this strategy, a training algorithm based on subtractive clustering is used to train the network. The radial basis function created by the newrb function from Matlab uses gradient based iterative method as the learning strategy. The proposed method is implemented in the Matlab which creates a new network that undergo a hybrid learning process. The networks called SC/RBF (Subtractive Clustering–Radial Basis Function) and SC/Modular RBF (Subtractive Clustering-Modular Radial Basis Function) are used to test against the standard Radial Basis Function and modular Radial Basis Function in protein classification. Classification criteria consist of two heuristic rules are implemented to test on the classification performance rate. The real world problem that has been considered is classification of human protein sequences into ten different superfamilies which based on protein function groups. These human protein sequences are downloaded from Protein Information Resource (PIR) database.
|Keyword:||Neural Networks, Protein Sequence Classification, n-gram method, bioinformatics|
|Faculty or Institute:||Institute for Mathematical Research|
|Deposited On:||09 Jun 2011 09:52|
|Last Modified:||27 May 2013 07:53|
Repository Staff Only: Edit item detail
Document Download Statistics
This item has been downloaded for since 09 Jun 2011 09:52.