Detecting "Pump & Dump Schemes" on Cryptocurrency Market Using An Improved Apriori Algorithm

With the popularity of Bitcoin, a cryptocurrency market emerged. However, because of insufficient supervision, the market attracts scams, for example, pump and dump (P&D) scheme, a famous fraudulent behavior in stock markets, has been found rampant in the market. To help deal with this issue, as a preliminary study, this paper proposes an improved apriori algorithm to detect user groups which may involve in P&D schemes. The validity of the algorithm is verified by using the leaked transaction history of Mt. Gox Bitcoin exchange. Furthermore, by exploring some of the detected user groups, many abnormal trading behaviors in the exchange found. These findings provide new insights into the behavior of users in the cryptocurrency market, thus leading to meaningful implications for policymakers, investors, and managers dealing with the cryptocurrency market.