Andro-Simnet: Android Malware Family Classification Dataset Using Social Network\u2013Based Similarity Measures

This dataset contains Android malware samples analyzed for constructing similarity-based malware family networks. Each sample was processed through both static and dynamic analysis to extract key behavioral and signature-based features. The extracted features include API call sequences, required permissions, activity names, and file names, which collectively capture functional and developer-centric characteristics of malware.Similarity values between malware pairs were computed using feature-specific measures\u2014Nilsimsa hashing for API call sequence similarity and Jaccard similarity for string-based features. These similarity values enable building a weighted malware similarity graph suitable for community-detection-based family classification. The dataset supports research involving malware clustering, behavioral analysis, social-network-based security analytics, and hybrid static\u2013dynamic malware characterization.This dataset is associated with the research described in \u201cAndro-Simnet: Android Malware Family Classification using Social Network Analysis\u201d presented at PST 2018. The paper provides details on feature extraction, similarity computation, weight optimization, and community detection techniques.