| Autor: |
Phogat*, Manu, Dharmender Kumar |
| Rok vydání: |
2021 |
| Předmět: |
|
| DOI: |
10.5281/zenodo.5720852 |
| Popis: |
Due to recent advancement in computational biology in past decade, the DNA microarray technology is appeared as a useful tool in detection of mutation in various complex diseases such as cancer. The availability of thousands of microarray datasets makes cancer classification an active research area. The gene microarray datasets suffer with curse of dimensionality problem. To reduce this problem and to enhance the informative gene selection process a new hybrid approach is proposed which combined improved binary competitive swarm optimization and whale optimization algorithm into IBCSOWOA. The proposed technique firstly used the minimum redundancy maximum relevance (mRMR) filter feature selection to identify the relevant gene subset from microarray datasets. After that wrapper method IBCSO is applied to reduce the informative gene subset originated from mRMR. The proposed approach use artificial neural network model with is parameter tuned by whale optimization algorithm (WOA), to classifying the disease accurately. The performance of proposed method is tested on six different mutation-based microarray datasets and compared with existing disease prediction methods. The experimental results show that proposed technique is significantly better than the existing nature inspired methods in terms of optimal feature subset, classification accuracy and convergence rate. The proposed technique reaches above 98% accuracy in in five datasets and highest accuracy achieved as 99.45 in Lung cancer dataset. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
