Optimizing cancer classification - a hybrid RDO - XGBoost approach for feature selection and predictive insights

Sep 21, 2026

Uploaded by Rowan

Identification of relevant biomarkers from high-dimensional cancer data remains difficult because cancer types are complex and heterogeneous. Conventional feature-selection methods often struggle to search the large solution space while preserving high predictive accuracy. A framework is proposed that integrates Random Drift Optimization (RDO) with XGBoost to improve cancer-classification performance. Experiments on real datasets for CNS, leukemia, breast, and ovarian cancers show improved accuracy and efficiency through selecting a smaller subset of unique, relevant genes. Compared with SVM, KNN, and Naive Bayes, the method consistently achieves superior accuracy and F-measure. Reported accuracies include 97.24% (CNS), 99.14% (leukemia), 95.21% (ovarian), and 87.62% (breast). These results suggest the RDO-XGBoost approach is promising for feature selection with strong predictive and biological insight.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages14

Reading Time35 min

AccessFree to download

Abstract

Keywords

Introduction

Cancer Immunology, Immunotherapy (2024) 73:261 [https://doi.org/10.1007/s00262-024-03843-x](https://doi.org/10.1007/s00262-024-03843-x) Optimizing cancer classification: a hybrid RDO‑XGBoost approach for feature selection and predictive insights Abrar Yaqoob1 · Navneet Kumar Verma1 · Rabia Musheer Aziz2 · Mohd Asif Shah3,4,5 Received: 28 August 2024 / Accepted: 20 September 2024 / Published online: 9 October 2024 © The Author(s) 2024 Abstract The identification of relevant biomarkers from high-dimensional cancer data remains a significant challenge due to the complexity and heterogeneity inherent in various cancer types. Conventional feature selection methods often struggle to effectively navigate the vast solution space while maintaining high predictive accuracy. In response to these challenges, we introduce a novel feature selection approach that integrates Random Drift Optimization (RDO) with XGBoost, specifically designed to enhance the performance of cancer classification tasks. Our proposed framework not only improves classification accuracy but also offers valuable insights into the underlying biological mechanisms driving cancer progression. Through comprehensive experiments conducted on real-world cancer datasets, including Central Nervous System (CNS), Leukemia, Breast, and Ovarian cancers, we demonstrate the efficacy of our method in identifying a smaller subset of unique and relevant genes. This selection results in significantly improved classification efficiency and accuracy. When compared with popular classifiers such as Support Vector Machine, K-Nearest Neighbor, and Naive Bayes, our approach consistently outperforms these models in terms of both accuracy and F-measure metrics. For instance, our framework achieved an accuracy of 97.24% in the CNS dataset, 99.14% in Leukemia, 95.21% in Ovarian, and 87.62% in Breast cancer, showcasing its robustness and effectiveness across different types of cancer data. These results underline the potential of our RDO-XGBoost framework as a promising solution for feature selection in cancer data analysis, offering enhanced predictive performance and valuable biological insights. Keywords Random drift optimization · XGBoost · Feature selection · Cancer classification · Microarray data analysis * Abrar Yaqoob [abraryaqoob77@gmail.com](abraryaqoob77@gmail.com) * Mohd Asif Shah [m.asif@kardan.edu.af](m.asif@kardan.edu.af) 1 VIT Bhopal University’s School of Advanced Science and Language, Located at Kothrikalan, Sehore, Bhopal 466114, India 2 Planning Department, State Planning Institute (New Division), Lucknow, Utter Pradesh 226001, India 3 Department of Economics, Kardan University, Parwane Du, 1001 Kabul, Afghanistan 4 Division of Research and Development, Lovely Professional University, Phagwara, Punjab 144001, India 5 Centre of Research Impact and Outcome, Chitkara University Institute of Engineering and Technology, Chitkara University, Rajpura, Punjab 140401, India Introduction Human genetic data have the potential to aid in the detection and categorization of a wide range of disorders, including cancer. Microarray analysis of gene expression levels in various individuals is one of the most precise technologies in this area. Each sample is represented by a row, and each gene or trait is represented in a column, making the microarray data a matrix with thousands of columns and a few hundred rows. Computational costs, the ability to generalize classifications, and their efficacy in predicting fresh microarray samples are all impacted by the huge number of genes and small number of samples [1, 2] . However, due to the large number of features, it is also possible that genes with no obvious connection to one another influence the development of predictive models. Since, biologically speaking, only a small subset of genes truly has a role in the disease, the remainder merely act as "background noise"that can obscure the effect of the first subset. Addi

FAQs

Why is feature selection challenging in cancer classification from microarray data?

High-dimensional cancer data create a large search space, and only a small subset of genes is biologically relevant. Irrelevant genes can act as background noise, reducing classification performance and increasing computational cost.

What does the proposed method combine, and what is its goal?

The framework integrates Random Drift Optimization (RDO) with XGBoost. Its goal is to improve cancer classification by selecting an informative subset of genes while maintaining high predictive accuracy.

How does the method perform compared with common classifiers?

The approach outperforms Support Vector Machine, K-Nearest Neighbor, and Naive Bayes in accuracy and F-measure metrics across multiple cancer datasets, including CNS, leukemia, breast, and ovarian cancers.

Sep 21, 2026

Uploaded by Rowan