Big Data Analysis is the market trend today in research institutes and industries. The data analysis problems can be solved by the machine learning and data mining techniques. The data mining techniques in the health care system are used to discover the valuable knowledge, which help the physicians to treat the patients at the earliest. This paper uses the data mining technology such as  feature selection and classification to develop a predictive model for ovarian cancer detection. A large data set is gathered and preprocessed. Rough set theory is used to discover the data dependencies and reduce the feature set contained in the dataset. The Hybrid PSO and ACO (PACO)is used to optimize the selected featuresto efficiently classify the ovarian cancer tumors, either malignant(Stage I, StageII, Stage III, Stage IV) or benign.The Classification task is performed by the Multi Layer Feed Forward Neural Network (MFFNN) and it is trained using the backpropagation algorithm with momentum. The performance of the system is measured in terms of classification accuracy, mean absolute error and root mean square error.