Early detection of lung cancer from chest CT images is a critical application of artificial intelligence in modern healthcare. This study proposes a method that integrates radiomics feature extraction with interpretable linear machine learning models to classify lung nodules as benign or malignant using the LUNA16 dataset. Following preprocessing and nodule localization based on provided coordinates, we employed the PyRadiomics library to extract 92 shape and texture features from the regions of interest (ROIs) containing nodules. Features were selected using recursive feature elimination (RFE) before training with a Ridge regression model. The dataset, comprising 856 lung nodules from 888 CT scans, was split into an 80/20 training/testing ratio. Experimental results demonstrate that the model achieved an accuracy of 87.5%, sensitivity of 83.1%, specificity of 89.9%, and an AUC of 0.91. The most significant features included gray-level intensity standard deviation, compactness, and GLCM entropy. The proposed method offers high accuracy, short training time, and strong interpretability, highlighting its potential for practical application in computer-aided medical imaging diagnosis systems.