ML Models for Cancer Prediction
Scrutinizing ML Models for Cancer Prediction
Built and evaluated machine learning models using patient records to predict lung cancer outcomes. Published findings in an academic journal.
Problem Statement
Cancer diagnosis requires high precision to minimize false negatives (missed diagnoses) and false positives (unnecessary treatments). While many ML models claim high accuracy, their performance varies significantly across different cancer types and patient demographics.
Methodology
Conducted a rigorous evaluation of multiple machine learning algorithms including Random Forest Classifier, K-Nearest Neighbors, Support Vector Machines, and K-Means clustering. Applied cross-validation and feature importance analysis on patient medical records to identify key predictive biomarkers.
Results
Identified that ensemble methods consistently outperform single models for cancer prediction. Published the comparative analysis and findings in the International Journal of Information Technology and Computer Engineering (IJITCE).