Revolutionizing Cervical Cancer Detection:SMOTE and Entropy Boosted Random Forest Approach

One of the most severe diseases that affect women is cervical cancer, which occurs when cells in the cervix or lower section of the uterus grow abnormally and develop into a cancerous tumor. The papillomavirus, sometimes known as HPV, is the primary cause of this particular type of cancer. If this kind of cancer is not treated, it will spread quickly to other body parts and cause grave complications. This cancer has two extreme features: it shows no symptoms in its early stages, there's a higher possibility of being treatable. The PAP test, which looks for abnormalities in uterine cells under a microscope and may detect the disease if there are any alterations, is the independent screening and testing procedure used to identify this form of cancer. For instance, this can occasionally result in a false positive, meaning the patient is diagnosed with cancer when in fact they do not have it, or a false negative, meaning that during sample collection, abnormal cells were either absent or improperly collected, or both. To mitigate these drawbacks, alternative approaches are needed to improve accuracy and enable early diagnosis, which will boost treatment effectiveness and survival rates. This can only be achieved through computerized methods, particularly with machine learning algorithms that allow for increased accuracy and the integration of multiple data sets, including electronic health records, PAP test results, and HPV test results. Therefore, machine learning model is created in this research study for the early prediction of cervical cancer using the Cervical Cancer dataset from the UCI Repository. There are 32 risk factors associated with each of the four target variables in this dataset: Himmelmann, Schiller, Cytology, and Biopsy. The dataset has a class imbalance issue, meaning that one set of records is significantly higher than another, leading to a biassed outcome. Therefore, the SMOTE approach is used in this research work to prevent biassed results and to improve accuracy. Metrics including Accuracy, Precision, Recall, and F1 Score are used to assess how well algorithms work in detecting cervical cancer