Android malware detection using PMCC heatmap and Fuzzy Unordered Rule Induction Algorithm (FURIA)

Abstract

Many smart mobile devices, including smartphones, smart televisions, smart watches, and smart vacuums, have been powered by Android devices. Therefore, mobile devices have become the prime target for malware attacks due to their rapid development and utilization. Many security practitioners have adopted different approaches to detect malware. However, its attacks continuously evolve and spread, and the number of attacks is still increasing. Hence, it is important to detect Android malware since it could expose a great threat to the users. However, in machine learning intelligence detection, too many insignificant features will decrease the percentage of the detection’s accuracy. Therefore, there is a need to discover the significant features in a minimal amount to assist with machine learning detection. Consequently, this study proposes the Pearson correlation coefficient (PMCC), a coefficient that measures the linear relationship between all features. Afterwards, this study adopts the heatmap method to visualize the PMCC value in the color of the heat version. For machine learning classification algorithms, we used a type of fuzzy logic called lattice reasoning. This experiment used real 3799 Android samples with 217 features and achieved the best accuracy rate of detection of more than 98% by using Unordered Fuzzy Rule Induction (FURIA).

Software Requirement Risk Prediction Using Enhanced Fuzzy Induction Models

“…Fuzzy rules are more generic than conventional rules, as they create dynamic boundaries for classification processes. For instance, the conventional rules work with fixed decision boundaries, with abrupt transitions between different classes, thereby generating questionable and non-intuitive models [21,22]. The understandability of the FURIA is, thus, a major advantage over the alternative (black box) conventional rule and ML-based methods, and a major reason for its adoption in this research.…”