Smartphones have become an important part of human lives, and this led to an increase number of smartphone users. However, this also attracts hackers to develop malicious applications especially Android botnet to steal the private information and causing financial losses. Due to the fast modifications in the technologies used by malicious application (app) developers, there is an urgent need for more advanced techniques for Android botnet detection. In this paper, a new approach for Android botnet classification based on features selection and classification algorithms is proposed. The proposed approach uses the permissions requested in the Android app as features, to differentiate between the Android botnet apps and benign apps. The Information Gain algorithm is used to select the most significant permissions, then the classification algorithms Na�ve Bayes, Random Forest and J48 used to classify the Android apps as botnet or benign apps. The experimental results show that Random Forest Algorithm achieved the highest detection accuracy of 94.6% with lowest false positive rate of 0.099. � 2017 American Scientific Publishers All rights reserved.

Along with the rapid development of new science and technology, smartphone functionality has become more attractive. Smartphones not only bring convenience to the public but also the security risks at the same time through the installation of malicious applications. Among these, Android ransomware is gaining momentum and there is a need for effective defense as it is very important to ensure the security of smartphone user. There are various analysis techniques used to detect instances of Android ransomware. In this paper, we proposed the Android ransomware detection using dynamic analysis technique. Two dataset were used which is ransomware and benign dataset. The proposed approach used the system calls as features which obtained from dynamic analysis. The classification algorithms Random Forest, J48, and Naïve Bayes were used to classify the instances based on the proposed features. The experimental results showed that the Random Forest Algorithm achieved the highest detection accuracy of 98.31% with lowest false positive rate of 0.016.

Nowadays mobile devices such as smartphones had widely been used. People use smartphones not limited for phone calling or sending messages but also for web browsing, social networking and online banking transaction. To certain extend, all confidential information are kept in their smartphone. As a result, smartphones became as one of the cyber-criminal main target especially through an installation of mobile botnet. Eurograbber is an example of mobile botnet that being installed via infected mobile application without victim knowledge. It will pretense as mobile banking application software and steal financial transaction information from victim's smartphone. In 2012, Eurograbber had caused a total loss of USD 47 Million accumulatively all over the world. Based on the implications posed by this botnet, this is the urge where this research comes in. This paper presents a proof of concept on how the botnet works and the ongoing research to detect and respond to the mobile botnet efficiently. Detection of botnet malicious activity is done through an analysis of Crusewind Botnet code using reverse engineering process and static analysis technique. � 2014 IEEE.

Android applications may pose risks to smartphone users. Most of the current security countermeasures for detecting dangerous apps show some weaknesses. In this paper, a risk assessment method is proposed to evaluate the risk level of Android apps in terms of confidentiality (privacy), integrity (financial) and availability (system). The proposed research performs mathematical analysis of an app and returns a single easy to understand evaluation of the app's risk level (i.e., Very Low, Low, Moderate, High, and Very High). These schemes have been tested on 2488 samples coming from Google Play and Android botnet dataset. The results show a good accuracy in both identifying the botnet apps and in terms of risk level.