Impact of PCA on Lung Cancer Dataset Classification: A Comparitive Analysis of Machine Learning Models

Cancer research relies on accurate classification models. Detecting cancer early enhances the chance of a cure. This paper proposes an innovative classification model using enhanced ML (Machine Learning) algorithms for early-stage Lung cancer diagnosis. The study compares the performance of Naive Bayesian, DT (Decision Tree), Ordinal LR (Logistic Regression), and SVM (Support Vector Machine) models before and after PCA (Principal Component Analysis) implementation. The primary objective is to assess how PCA influences classification accuracy in cancer datasets. This study initially evaluates the accuracy rates of four models on a Lung cancer dataset of 250 instances, showing a decrease in accuracy for all four models after PCA implementation, with the Decision Tree model achieving a positive accuracy rate of 94 percent. To enhance accuracy, the dataset volume was increased to 1000 instances, which showed that the Decision Tree model consistently performed exceptionally well, reaching a 100 percent accuracy rate. This highlights the robustness of the DT model and the critical role of model selection in Lung cancer Research.