Published Paper Details:

Colorectal cancer is still a major global public health concern that requires accurate and timely identification to intervene effectively. This study uses a large dataset with clinical and pathological variables to investigate the use of logistic regression for colon cancer identification. The dataset includes demographic information, lesion characteristics, and histological results obtained from colonoscopy tests. The dataset was preprocessed using feature engineering approaches, which included fixing missing values and normalizing variables to ensure the best possible model performance. Known for its effectiveness in binary classification tasks, logistic regression was the main predictive model. After thorough training and assessment, the logistic regression model showed encouraging results in differentiating between benign and malignant colon tumors. Evaluation metrics included the receiver operating characteristic (ROC) curve, recall, accuracy, and precision.