A heuristics approach to mine behavioural data logs in mobile malware detection system

Abstract

Nowadays, in the era of every "things" connect together via the Internet, mobile device number has increased exponentially up to tens billions around the world. In line with this equipment increase, generated data amount is enormous and have attracted malefactors to spoil. For hackers, one of the popular ways to threaten mobile devices is to spread malware, which is very difficult to prevent because the application installation and configuration rights are set by owners, who usually have very low knowledge or do not care about security. In this study, we aim at improving security in environment of mobile devices by proposing a novel system to detect automatically malware intrusions. Our solution thus is built based on modeling user behaviors and using heuristic analysis approach to mobile logs generated during the device operation process. Although behaviors of individual user have a significant impact upon social cyber-security but achievement of user awareness still remains one of the major challenges today. For this task, a light-weight semantic formalization in form of physical and logical taxonomy is proposed for classifying the collected raw log data. Then we use a set of techniques like sliding windows, lemmatization, feature selection, term weighting and so on to process data. Meanwhile, malware detection tasks are performed based on incremental machine learning mechanisms because tasks' complex degree. Our solution is developed in the manner of allowing scalability in which the system is composed by several component blocks that cover preprocessing raw collected logs from mobile devices, automatically creating datasets for machine learning methods, using the best selected model for detection suspicious activity surrounding malware intrusions and supporting decision making using predictive risk. The proposals are experimented cautiously and gained test results show the effectiveness and feasibility of our malware detection system in applying to large-scale mobile environment.

Machine Learning and Deep Learning frameworks and libraries for large-scale data mining: a survey

“…At the first sight, general audience relates Big Data processing with distributed platforms like Apache Hadoop and Spark. The contiguity goes there for Volume in conformity with Veracity characteristic (Nguyen et al 2018) as well as high-speeds in processing and inference in background. In the new era of Big Data, data analysis is expected to change.…”

Hybridization of Galactic Swarm and Evolution Whale Optimization for Global Search Problem

“…Heuristic algorithms are often used for optimization problems based on several assumptions, such as knowledge inside a concrete domain [19]. Meta-heuristic algorithms are more domain-independent, and often operate based on probabilistic rules rather than deterministic ones.…”

Real-time data exploitation supported by model- and event-driven architecture to enhance situation awareness, application to crisis management

“…Among them seven were proposing a general approach, or an architecture, to ascend the three levels of situation awareness, but none of them was tackling more than the volume of Big Data. This concerns Avvenuti et al (2018), who analyse social media data; Wu et al (2018), who propose a security situational awareness; Nguyen et al (2018), who detect malware intrusions automatically; Hingant et al (2018), who propose to critical important inherited from the perception step Figure 2. The proposed framework to support the design of a new emergency decision support system, able to connect to diverse, external, data sources, in order to improve their users' situation awareness.…”