Machine learning techniques have been extensively applied to help medical experts in making a diagnosis of many diseases. Classification is a machine learning technique that is used to forecast the relationship between data samples and classes. It is an essential task in different applications, such as image classification and medical diagnosis. There are different classification techniques, such as SVM, C5.0, Neural Network, K-Nearest Neighbor, and Naive Bayes Classifier. Feature selection for classification of cancer data means discovering feature values of malignant tumors and benign ones. It also means using this knowledge to forecast the state of new cases. In this paper, we use Rough sets as a feature selection technique to create a subset feature from the original features. Therefore, we use the resulting subset with different classification and ensemble techniques to discover classes of unknown data using HCV data set. SVM, C5.0, and Ensemble classifiers are used as classification techniques to discover classes of unknown data. In this paper, the percentage of accuracy, sensitivity, and specificity are used as evaluation parameters for the tested classification techniques. Experimental results show that the proposed hybrid RS-Boosting/SVM technique has higher accuracy, sensitivity and specificity rates with selected subset features than other tested techniques.