An Advance Implementation of Machine Learning Techniques for the Prediction of Cervical Cancer

Abstract

Cancer has now become a widespread issue, and it might be difficult to diagnose it. The analysis and diagnosis of disease in a patient through a monitoring system has advanced significantly, but there are still some obstacles to overcome. Technology-assisted disease discovery is urgently required since it could enable clinicians to make an accurate diagnosis more quickly. The process makes use of event-driven datasets produced by the Internet of Things (IoT). Rarely are the cloud and IoT used to process the real-time dataset for illness discovery. Cancer detection is a serious difficulty since the data generated by sensors constitutes huge data and cannot be quickly processed without the use of data mining and machine learning. A precise and intelligent method is needed for cancer prognosis because it has been discovered that cancer cannot be accurately predicted from offline datasets. The International Federation of Obstetrics and Gynecology (IFOG) realworld data set has been used to predict cervical cancer using machine learning. Five metrics, including true-positive-rate, false-positiverate, f-measure, MCC, and accuracy, have been used to examine the results of the training of six machine learning models using the latest UCI repository. The outcomes demonstrate that the tree-based decision stump machine learning classifier and the meta-based iterative classifier optimizer achieved the maximum prediction accuracy of 77.97%, which is the greatest among the other ML classifiers. The proposed model demonstrated noteworthy performance in terms of prediction accuracy.

Topics & Concepts