An Enhanced Prediction of Ovarian Cancer based on Ensemble Classifier using Explainable AI

For women's general health and well-being, ovarian cancer detection and prevention are of paramount importance. Opposite uterine cancer, sometimes known as the "silent killer," presents with subtle signs and symptoms in the early stages, making it difficult to detect in a timely manner. The chance of successful treatment and survival from ovarian cancer is greatly reduced when caught in its later stages. In order to catch the disease in its early, more curable phases, it is crucial to conduct screenings on a regular basis. These screenings can include pelvic exams, ultrasounds, and blood testing for specific biomarkers. This study employs the ovarian cancer dataset from Soochow University, which includes 50 features for precise cancer identification. In order to improve the accuracy and dependability of predictions, the suggested model employs a stacked ensemble-model, which combines the advantages of bagging &boosting classifiers. Better ovarian cancer prediction results are a result of this combination's use of variance reduction and enhanced generalization. With all features considered, the suggested model achieves the greatest model result to date on this dataset, with an accuracy of 96.87%. The outcomes are further explained utilizing SHAPly, an explainable-AI approach. The proposed model's superiority is proven by contrasting its results with those of another state-of-the-art model.