A Comparative Study on Machine Learning Classifiers for Early Diagnosis of Cervical Cancer

Cervical cancer is a significant worldwide health issue, particularly in underdeveloped countries where the availability of preventive healthcare treatments is restricted. Effective prevention and treatment of cervical cancer depend on early recognition of the disease's threats. Machine-learning algorithms have become popular in recent years as promising methods of diagnosing cancer risk based on medical and demographic data. This study looks to observe the way algorithms based on machine learning might be applied to estimate the risk of cervical cancer with a dataset that includes patient demographics, clinical history, and diagnostic test findings. Predictive systems are created using several machine learning approaches, such as Decision Tree, Naïve Bayes, Support Vector Machine, K-Nearest Neighbors, Random Forest, Logistic Regression, Gradient Boosting, Nearest Centroid, Multilayer Perceptron, and AdaBoost. These model's prediction powers are evaluated using performance criteria including accuracy, sensitivity, precision, f-measure, specificity, and area under the receiver operating characteristic curve (AUC-ROC). Our findings indicate that the decision tree achieved the maximum accuracy, precision, and f1-score (98.91%, 97.81%, and 0.9889, respectively). Furthermore, hyperparameter tuning was used to optimize model performance. For instance, the Support Vector Machine (SVM) attained a remarkable accuracy of 99.27%, precision of 98.53%, and F1-score following hyperparameter adjustment of 0.9926, demonstrating its potential in cervical cancer chance prediction. This demonstrates the possibilities and capability of machine learning methods in improving correct predicting as well as patient outcomes for cervical cancer detection.