Enhancing Cancerous Gene Selection and Classification for High-Dimensional Microarray Data Using a Novel Hybrid Filter and Differential Evolutionary Feature Selection

Sep 21, 2026

Microarray datasets support early-stage cancer diagnosis by representing gene expression patterns, yet they typically include thousands of redundant, irrelevant, and noisy genes that make machine learning on high-dimensional data difficult. The study presents a two-phase hybrid feature selection framework combining popular filter methods with differential evolution (DE). Filter scoring retains top-ranked genes to reduce dimensionality, and DE further refines the optimal subset. Four cancerous microarray datasets (Breast, Lung, CNS, Brain) validate the approach, yielding markedly improved classification accuracy up to 100% and notable feature reduction.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages37

Reading Time1 h 33 min

AccessFree to download

Introduction

Background of microarray-based cancer analysis

Methods

Two-phase hybrid filter-DE feature selection

Differential evolution optimization

Experiments

Datasets used for evaluation

Classification performance and accuracy improvements

Results and Discussion

Accuracy gains over filter-only baselines

Feature reduction compared with previous work

Conclusions

Article Enhancing Cancerous Gene Selection and Classification for High-Dimensional Microarray Data Using a Novel Hybrid Filter and Differential Evolutionary Feature Selection Arshad Hashmi 1, *, Waleed Ali 2, *, Anas Abulfaraj 1, Faisal Binzagr 3 and Entisar Alkayal 2 Citation: Hashmi, A.; Ali, W.; Abulfaraj, A.; Binzagr, F.; Alkayal, E. Enhancing Cancerous Gene Selection and Classification for HighDimensional Microarray Data Using a Novel Hybrid Filter and Differential Evolutionary Feature Selection. Cancers 2024, 16, 3913. [https://](https://)[ ](https://)[doi.org/10.3390/cancers16233913](doi.org/10.3390/cancers16233913) Academic Editor: Jason Roszik Received: 18 September 2024 Revised: 6 November 2024 Accepted: 19 November 2024 Published: 22 November 2024 Copyright: © 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . 1 Department of Information Systems, Faculty of Computing and Information Technology, King Abdulaziz University, P.O. Box 344, Rabigh 21911, Saudi Arabia 2 Information Technology Department, Faculty of Computing and Information Technology, King Abdulaziz University, P.O. Box 344, Rabigh 21911, Saudi Arabia; [ealkayyal@kau.edu.sa](ealkayyal@kau.edu.sa) 3 Department of Computer Science, Faculty of Computing and Information Technology, King Abdulaziz University, P.O. Box 344, Rabigh 21911, Saudi Arabia; [fbinzagr@kau.edu.sa](fbinzagr@kau.edu.sa) * Correspondence: [ahsyed@kau.edu.sa](ahsyed@kau.edu.sa) (A.H.); [waabdullah@kau.edu.sa](waabdullah@kau.edu.sa) (W.A.) Simple Summary: To improve cancer classification performance for high-dimensional microarray datasets, this work proposes combining filter and differential evolutionary (DE) algorithm featureselection techniques. By scoring genes or features of high-dimensional microarray datasets by some common filter methods, we keep only the highest-ranked features and eliminate superfluous and irrelevant ones to decrease the dimensionality of the microarray datasets. Then, the genes or features of the microarray datasets are optimized further by DE, producing noticeably better classification results. This could lead to outstanding improvement in the cancer classification using only less features of the microarray datasets. Abstract: Background: In recent years, microarray datasets have been used to store information about human genes and methods used to express the genes in order to successfully diagnose cancer disease in the early stages. However, most of the microarray datasets typically contain thousands of redundant, irrelevant, and noisy genes, which raises a great challenge for effectively applying the machine learning algorithms to these high-dimensional microarray datasets. Methods: To address this challenge, this paper introduces a proposed hybrid filter and differential evolution-based featureselection to choose only the most influential genes or features of high-dimensional microarray datasets to improve cancer diagnoses and classification. The proposed approach is a two-phase hybrid feature selection model constructed using selecting the top-ranked features by some popular filter feature selection methods and then further identifying the most optimal features conducted by differential evolution (DE) optimization. Accordingly, some popular machine learning algorithms are trained using the final training microarray datasets with only the best features in order to produce outstanding cancer classification results. Four high-dimensional cancerous microarray datasets were used in this study to evaluate the proposed method, which are Breast, Lung, Central Nervous System (CNS), and Brain cancer datasets. Results: The experimental results demon

FAQs

Why is gene selection critical for high-dimensional microarray cancer datasets?

Microarray datasets contain thousands of redundant, irrelevant, and noisy genes, which increases dimensionality and makes machine learning harder to apply effectively. Selecting influential features reduces noise and improves classification quality.

How does the proposed hybrid method work in two phases?

First, filter-based feature selection scores and keeps the highest-ranked genes to reduce dimensionality. Second, differential evolution (DE) optimizes the remaining candidates to identify an even more optimal gene subset.

What performance improvements are reported on the evaluated cancer datasets?

The hybrid filter-DE approach reaches accuracy values of 100%, 100%, 93%, and 98% for Brain, CNS, Breast, and Lung datasets. It also removes around 50% of features selected by filter methods.

What is the main takeaway compared with previous feature selection works?

The method shows better improvement percentages for Brain and CNS, comparable gains for Lung, and smaller improvement for Breast when compared with prior approaches.

Sep 21, 2026

Uploaded by Clementine