High Performance Cervical Cancer Detection Using Gradient Boosting Algorithm and Bayesian Optimization
M. Sangeetha, Vijayan Sugumaran, T. Karthick, Jagadeesh Puli
Abstract
Early treatment, a higher survival rate, fewer fatalities, and a higher standard of living for individuals who have been diagnosed with the disease are all benefits of cervical cancer screening. It is an important field for study and development because precancerous lesion treatment can help stop the progression of the cancer and stop the onset of cervical cancer. Based on the dataset, this paper provides a method for detecting cervical cancer using a gradient boosting technique. Selected characteristics and preprocessed data were used to train the model. Bayesian optimization was used for hyperparameter tuning to determine the ideal hyperparameter values. Using evaluation criteria such as precision, accuracy, F1-score, recall and AUC-ROC the effectiveness of the model was assessed. The findings demonstrate that the suggested method performed well in detecting cervical cancer, with precision, accuracy, F1-score and recall all exceeding 88%. In addition, the area under the ROC curve was 92% (AUC-ROC). The results show that utilizing the gradient boosting algorithm in conjunction with Bayesian optimization for hyperparameter tweaking on the dataset can be a successful method for detecting cervical cancer. fostering community, dialogue, and outcomes in the fields of leadership development, mentoring, entrepreneurship, and social good in order to support the professional achievement of South Asian women.