Class imbalance problem are raised when one class having maximum number of examples than other classes. The classical classifiers of balance datasets cannot deal with the class imbalance problem because they pay more attention to the majority class. The main drawback associated with it majority class is loss of important information. The Class imbalance problem is a difficult due to the amount and nature of data. This paper focuses different methods of class imbalance problem. It is been considered the majority class to achieve the class imbalanced problem. This paper mainly focuses the minority class sample to achieve the problem and proposed method for class imbalance problem using minority sample data. The oversampling and undersampling both concept were used to identify the correct class label of the sample using probabilistic approach, the main objective of this paper, to proposed method to minimize the misclassification rate of minority class sample, balance and classify the data more accurately thereby improving the performance of classifier.