Unsupervised Learning Based Detection Method for The Life Cycle of Virtual Cryptocurrency

Abstract

Since the birth and open source of Bitcoin, there are more than 10,000 kinds of virtual cryptocurrencies in the market. Every day, virtual cryptocurrencies are born, but also virtual cryptocurrencies die out. In the life cycle of virtual cryptocurrencies, different periods of abnormal transactions will occur. However, there are still deficiencies in the definition and related studies of the life cycle of virtual cryptocurrencies in existing researches. Machine learning (ML) can dig out the hidden rules from a large amount of data. We use unsupervised learning in machine learning to detect the life cycle of virtual cryptocurrencies. In this work, we divide and define each stage of the life cycle of virtual cryptocurrencies in detail. Based on the popularity value system of virtual cryptocurrencies and the similarity comparison algorithm, we establish a virtual cryptocurrencies life cycle detection tool to detect the life cycle stage of virtual cryptocurrencies. Experimental results demonstrate the effectiveness of the proposed unsupervised learning based virtual cryptocurrency life cycle detection tool.