A Machine Learning Framework for Integrating Multi-Omics Data for Early Leukemia Detection

The early detection of leukemia remains a critical challenge in medical diagnostics due to the complexity and variability of omics data. This research introduces a novel machine learning framework that integrates multi-omics data for the early detection of leukemia, utilizing the proposed CCA-RFE Selector (CCARS) algorithm. The CCARS algorithm combines Canonical Correlation Analysis (CCA) with Recursive Feature Elimination (RFE) to enhance feature selection and classification accuracy. The effectiveness of CCARS is evaluated through extensive simulation analyses, comparing its performance with existing algorithms, including Principal Component Analysis (PCA), Lasso Regression (LR), and Canonical Correlation Analysis (CCA). The evaluation metrics used include accuracy, F1-score, and computational time. Results demonstrate that the CCARS algorithm significantly improves accuracy and F1-score while reducing computational time compared to the existing methods. The CCARS algorithm achieves notable performance enhancements, validating its potential for integrating multi-omics data in leukemia detection.