INTEGRATED HYBRID MACHINE LEARNING ALGORITHMIC APPROACH FOR LUNG CANCER PREDICTION BASED ON CHARACTERISTICS FEATURES
TL;DR
This work points out the LR, DT, and KNN as the best models that achieve a high accuracy in their classification, promising candidates for real-world application for the prognostication of lung cancer.
Abstract
Lung cancer currently is amongst the most lethal virulence’s in the world currently, where late diagnosis is fundamentally responsible for the poor survival rates. Although it goes without saying, early detection will bring improvement to treatment outcomes, in that the process is really challenged by the restricted capabilities of the current imaging tools and symptoms overlap. Our approach includes machine learning algorithms such as Logistic Regression, Support Vector Machine, Decision Tree, K-Nearest Neighbors and Naive Bayes, for the classification of lung cancer cases based on patient data. So, the potential of all these methods in identifying lung cancer effectively is estimated by analyzing the performance of each of them using performance metrics such as the precision, accuracy, recall, and F1-score. The dataset, sourced from Manipal Hospitals, Bangalore, consists of multiple factors, including demographic, environmental, and lifestyle attributes along with health and genetic factors. Such a finding points out the LR, DT, and KNN as the best models that achieve a high accuracy in their classification, promising candidates for real-world application for the prognostication of lung cancer.