A Machine Learning Framework Developed by Leveraging the Topological Pattern Analysis of Bitcoin Addresses in Blockchain Network for Ransomware Identification

The steep uprise of the incognito cryptocurrency transactions has significantly raised the chances for the ransomware developers of demanding ransom by enciphering sensitive data of not only the individuals but also the large corporate houses. Majority of the recent ransomware operators prefer bitcoin to be a medium of their murky transactions. Although these bitcoin postings are permanently documented in the blockchain ledger, existing practices to detect & mine the origin of these ransomwares are still very difficult, which consist of numerous data gathering steps – making the process much lengthy & complex. As a solution to that, multiple statistical tools machine learning algorithms like Decision Tree, Random Forest, Binomial & Multinomial Logistic Regression has been utilized and a framework has been suggested to detect malicious addresses automatically with a considerably low lead time, by capturing the input data from coin movement topology while leveraging the graph theory & network analysis methodologies as its foundation.