Android developer develops applications and stores it in Google play market to earn profits. Some developers may repackage other's application (Free apps available in Android market) with their names or include various advertisements and add some malicious code to the apps and upload them to the application market for their profit. Similarly, attackers may also include malicious code into the original application in order to gain access the mobile devices to retrieve the personal information of the user, send SMS messages to premium rate numbers etc. Many detection methods have been used to find these repackaged applications available in various markets but they are not so effective to analyze the unknown applications. An efficient method is essential to reduce the number of features, decrease the classification error and the processing time. Repacked apps are identified by comparing the permissions of original and repacked apps. In this paper, a method is proposed to detect a repackaged application by using mutual information and chi square for selecting features and evaluating using random forest classifier. The performance of the classifier is compared with SVM, NB, Decision Tree and KNN.