A comprehensive machine learning technique for Ovarian Cancer prediction

Abstract

On the list of most prevalent cancers in women, ovarian cancer comes in eighth place. Because of standard screening and surveillance limitations, it is clinically impractical to analyse tumour molecular markers to predict therapy response. A relevant dataset with the necessary features was chosen for the prediction task of whether the patient has ovarian cancer or not. This approach uses Multilayer perceptron, ELM with AdaBoost, XGboost, LSTM, and a new CNN with Random Forest algorithms to predict Ovarian Cancer. The assessment parameters for each model were calculated for accuracy, precision, recall, F1-Score, Jaccard Index, and Error rate. The algorithms are then compared based on these metrics to determine the best algorithm. The CNN with Random Forest was the best method, with 95 to 100% accuracy. The CNN with Random Forest algorithm outperformed the individual existing ensemble techniques studied.