Comparative Study of Machine Learning and Deep Learning Models for Early Prediction of Ovarian Cancer
Hardik Dhingra, Roopashri Shetty · IEEE Access · 2025
Ovarian cancer is one of the most challenging cancers to detect early, often leading to poor survival rates. This study explores machine learning and deep learning approaches to improve predictive accuracy using clinical and biomarker-based data. The research begin by carefully preprocessing the dataset by handling missing values, removing outliers, applying principal component analysis (PCA), and normalizing data to ensure high-quality inputs. Various classifiers, including k-Nearest Neighbors (KNN), Support Vector Machine (SVM), Logistic Regression (LR), and Random Forest (RF), are tested, while feature selection techniques such as Feature Importance, Recursive Feature Elimination (RFE), Univariate Selection, and Correlation Analysis help refine input variables for better model efficiency. To further boost performance, ensemble methods like Stacking, Bagging, XGBoost, and AdaBoost are incorporated. Additionally, Deep Learning models such as Artificial Neural Networks (ANN), Feedforward Neural Networks (FNN), Recurrent Neural Networks (RNN), and Convolutional Neural Networks (CNN) are evaluated, with an autoencoder-based feature selection approach optimizing results. The effects of different dimensionality reduction techniques and the impact of standardization versus normalization are compared. Our findings highlight that a well-optimized combination of feature selection, ensemble learning, and deep learning significantly enhances ovarian cancer prediction, providing a valuable foundation for early diagnosis and clinical decision support.