Methodology Interaction by Machine Learning Model to Detect Vulnerability in Smart Contract of Blockchain

Abstract

Smart contracts have become increasingly popular in the development of trustworthy decentralized applications in recent years. These tools compare vulnerable contracts to a set of predefined rules. However, the emergence of new vulnerable types and programming skills to mitigate potential vulnerabilities results in many false positive and false negative tool reports. To address this, this data was analyzed using unsupervised machine learning to determine whether an algorithm can distinguish between shady/illegal owners and clean owners. Clustering algorithms are used in this paper. However, algorithms can cluster objects and detect fraud activity in Bitcoin transactions. Research on bitcoin network anomalies and suspicious transactions seeks to identify anomalous transactions, when all nodes on the bitcoin network are unlabeled. There is no evidence that any transaction is illegal. We are primarily interested in discovering irregularities in the bitcoin transaction network.