Int. J. Computer Applications in Technology, Vol. 59, No. 2, 2019 165
Copyright © 2019 Inderscience Enterprises Ltd.
A hybridised feature selection approach in
molecular classification using CSO and GA
Ahmed Elsawy, Mazen M. Selim and
Mahmoud Sobhy*
Computer Science Department,
Faculty of Computers and Informatics,
Benha University,
Benha, Qalyubia Governorate, Egypt
Email: ahmed.el_sawy@fci.bu.edu.eg
Email: selimm@fci.bu.edu.eg
Email: mahmoud.hassan@fci.bu.edu.eg
*Corresponding author
Abstract: Feature selection in molecular classification is a basic area of research in
chemoinformatics field. This paper introduces a hybrid approach that investigates the
performances of chicken swarm optimisation (CSO) algorithm with genetic algorithms (GA) for
feature selection and support vector machine (SVM) for classification. The purpose of this paper
is to test the effect of elimination of the inconsequential and redundant features in chemical
datasets to realise the success of the classification. The proposed algorithm was applied to four
chemical datasets and proved superiority in achieving minimum classification error rate in
comparison with different feature selection algorithms for molecular classification.
Keywords: molecular classification; chicken swarm optimisation; genetic algorithms; support
vector machines; feature selection.
Reference to this paper should be made as follows: Elsawy, A., Selim, M.M. and Sobhy, M.
(2019) ‘A hybridised feature selection approach in molecular classification using CSO and GA’,
Int. J. Computer Applications in Technology, Vol. 59, No. 2, pp.165–174.
Biographical notes: Ahmed Elsawy is Head of Computer Science Department in Faculty of
Computers and Informatics at Benha University, Egypt. He is Associate Professor in Computer
Science since 2017. He received PhD in 2011 in artificial intelligence and optimisation. He is
interested in how to use artificial intelligence in optimisation, bioinformatics, wireless network
and image recognition problems.
Mazen M. Selim received the BSc in Electrical Engineering in 1982, the MSc in 1987 and PhD in
1993 from Zagazig University (Benha Branch) in electrical and communication engineering. He
is Associate Professor in Computer Science Department at the Faculty of Computers and
Informatics, Benha University. He is currently the Vice Dean of Students and Learning Affairs at
Faculty of Computers and Informatics, Benha University. His research interests are in image
processing, biometrics, sign language, content-based image retrieval (CBIR), face recognition
and watermarking.
Mahmoud Sobhy received BSc degree in Computer Science in 2011 from Faculty of Computers and
Informatics, Benha University, and MSc degree in Applying Bio-inspired Algorithms in Molecular
Classification in 2017. Now, he works as a Lecturer Assistant in Computer Science Department in
the Faculty of Computers and Informatics at Benha University, Egypt. His research interests are in
using swarm intelligence algorithms for chemoinformatics and bioinformatics.
1 Introduction
Developing a new chemical entity as a drug is still a
challenging, time-consuming and cost-intensive process
(Terfloth and Gasteiger, 2001). Decreasing costs and speeding
up the discovery process are the primary objectives of drug
discovery, both in pharma and biotech sector (Pereira
et al., 2013). Improvements in computational techniques
suggest an alternative to medical chemistry techniques for
studying the structure and foretelling the biological activity of
drug candidates and that way highly minimising classical
resource requirements (Huang et al., 2010).
One of the key problems in the structure-activity
relationship research focuses on the characterisation and
quantification of chemical structures, as an appropriate
correlation can only be developed if both the biological activity