Hybrid Modelling of Pulmonary Cancer Risk Prediction Using Classical Algorithms to Modern Machine Learning Techniques

Description

Seeta Devi, Roshan Yadav, Joyce Robert Mathivanan, Sonopant Joshi, and Bhagyashree Jogdeo tested hybrid machine learning models for early lung cancer prediction in twenty twenty-six. Late diagnosis makes this cancer a leading cause of cancer death, so data-driven prediction is a live goal. Their data set held one thousand patient records with twenty-five candidate factors drawn from medical records.

After cleaning and trimming, eighteen factors stood out, spanning symptoms, lifestyle, and exposures. Age, passive smoking, obesity, smoking, and coughing blood all tracked the disease strongly. Many models, from decision trees to random forests, hit perfect scores on the test metrics, while deep learning and bagged models scored near ninety-seven percent. The authors say machine learning, above all support vector machines, holds strong promise for early prediction from structured records.

Catalog record

Source: https://europepmc.org/articles/PMC13651436