User:Nazrin93/sandbox

Article : (DyHAP) Dynamic Hybrid ANFIS-PSO Approach for Predicting Mobile Malware

Citation: Afifi, F., Anuar, N. B., Shamshirband, S., & Choo, K. K. R. (2016). DyHAP: Dynamic Hybrid ANFIS- PSO approach for predicting mobile malware. PLoS ONE, 11(9), 1–21.

Research Paradigm: Positivism – In the research is more on quantitative approach, where their study is to identify and optimum the model/parameter that used to detect mobile malware. Thus, to prove their proposed methods they need to measure the performance of the algorithm that been optimized.

Purpose of Study: The main purpose of research to identify the optimum parameters that can be used to facilitate mobile malware identification.

Research Objectives/ Research Questions: To optimum parameters that use for mobile malware detection. While the research questions are as follows;

What is the algorithms can be used to do optimization? How is the performance of prediction after optimization? Research Method: In the research at beginning the paper mention that their proposed method have two phase where are data collection & feature selection and extraction & labelling phase.

Sample used: They gathered and analyzed network traffic developed by android apps. The running apps will capture the network traffic in real-time network environment every 30 minutes. There manage to captured 1000 samples from 1260 malware data samples from 49 families in the Malgenome dataset as their references. Then, the samples were analyzed with public malware detection sandbox (Anubis Iseclab) and automatic android program analysis (SanDroid).

Method Proposed: Their approach method in the research is combines adaptive neuro fuzzy inference system (ANFIS) and particle swarm optimization (PSO). By using PSO it manage to improve performance of ANFIS which it adjust the membership function and minimizing error.

Main Findings: The research focus on network based features where application traffic was filtered for its parameters and calculation was performed on these parameters to obtain the required features. They also proposed three system agents to capture and manage the pcap file for the data preparation phase that improve on the efficient data processing. Then, they proposed a hybrid methods where are combination an adaptive neuro fuzzy inference system (ANFIS) and particle swarm optimization (PSO) to optimize the parameters that used in mobile malware detection.

Limitations of study: The study only show the performance of algorithms when the proposed method optimize the parameters but not discussing on the accuracy of detection whether get better true positive detection or vice versa.

Suggestions for future research: It is clearly state that to refined selection of variable it is because of evolution of malware and selection of input variables, such as identifying and discarding irrelevant variables.

Reference:

https://doi.org/10.1371/journal.pone.0162627