Prominent Binary-Feature (Permissions) Frequencies for Android Mobile Benign Apps and Malware Datasets

Published: 4 September 2020| Version 1 | DOI: 10.17632/ptd9fnsrtr.1
Gürol Canbek


The following academic datasets provided the frequencies of prominent binary-features, namely Android application permission requests* for the negative and positive classes. They are used in analyzing the dataset dissimilarities via the method proposed by Gürol Canbek. Negative/Positive dataset pairs** and their references - DS0 by (Lindorfer et al., 2014). (ANDRUBIS) as it is called "Touchstone Dataset" in the related article - DS1 by (Aswini and Vinod, 2014). (Contagio) - DS2 by (Wang et al., 2014)*** - DS3 by (Yerima et al., 2014)*** - DS4 by (Jiang and Zhou, 2013). (Android Malware Genome Project, AMGP) - DS5 (Peng et al., 2012) Thanks to the authors providing the data. * ** DS4 is a positive-class only dataset (it does not negative-class samples) *** DS2 and DS3 positive-class datasets contain DS4 (AMGP) samples.



Cybersecurity, Mobile Computing, Machine Learning, Feature Selection, Mobile Device, Malware Mitigation, Similarity Measure, Data Quality, Multivariate Quantitative Analysis, Performance Evaluation, Classifier Evaluation, Kruskal-Wallis H Test