Cryptocurrency ecosystems face increasing challenges associated with fraudulent activities, necessitating sophisticated solutions for timely detection and prevention. The heuristic and mark based approaches were the underpinning of before location strategies, yet unfortunately, these techniques were lacking to investigate the whole intricacy of peculiarity identification. With this contextual understanding, isn’t it true that prevention (avoiding sending crypto to scammers) is a better option than cure (retrieving sent crypto). The proposed system integrates multiple machine learning models into an ensemble, harnessing the collective intelligence of diverse algorithms for improved accuracy and robustness in identifying fraudulent patterns. Experimental evaluations, conducted on historical data and simulated scenarios, demonstrate the effectiveness of the proposed ensemble learning framework in bolstering the security and trustworthiness of crypto currency transactions. This is why the aim of this project is to develop an application whereby users can check whether an address is a potential fraudulent one before making the irreversible decision of sending crypto over to that address. This will be done by developing machine learning models (Ensemble learning) which use common attributes of crypto addresses to make calculated decisions on whether an address might be potentially fraudulent.
Keywords—Ensemble, Machine Learning, Machine Learning Algorithms, crypto currency transaction.