Financial price bubbles have previously been linked with the epidemic-like spread of an investment idea; such bubbles are commonly seen in cryptocurrency prices. This paper aims to predict such bubbles for a number of cryptocurrencies using a hidden Markov model previously utilised to detect influenza epidemic outbreaks, based in this case on the behaviour of novel online social media indicators. To validate the methodology further, a trading strategy is built and tested on historical data. The resulting trading strategy outperforms a buy and hold strategy. The work demonstrates both the broader utility of epidemic-detecting hidden Markov models in the identification of bubble-like behaviour in time series, and that social media can provide valuable predictive information pertaining to cryptocurrency price movements.

Cryptocurrencies have experienced recent surges in interest and price. It has been discovered that there are time intervals where cryptocurrency prices and certain online and social media factors appear related. In addition it has been noted that cryptocurrencies are prone to experience intervals of bubble-like price growth. The hypothesis investigated here is that relationships between online factors and price are dependent on market regime. In this paper, wavelet coherence is used to study co-movement between a cryptocurrency price and its related factors, for a number of examples. This is used alongside a well-known test for financial asset bubbles to explore whether relationships change dependent on regime. The primary finding of this work is that medium-term positive correlations between online factors and price strengthen significantly during bubble-like regimes of the price series; this explains why these relationships have previously been seen to appear and disappear over time. A secondary finding is that short-term relationships between the chosen factors and price appear to be caused by particular market events (such as hacks / security breaches), and are not consistent from one time interval to another in the effect of the factor upon the price. In addition, for the first time, wavelet coherence is used to explore the relationships between different cryptocurrencies.

Cryptocurrencies have recently experienced a new wave of price volatility and interest; activity within social media communities relating to cryptocurrencies has increased significantly.There is currently limited documented knowledge of factors which could indicate future price movements. This paper aims to decipher relationships between cryptocurrency price changes and topic discussion on social media to provide, among other things, an understanding of which topics are indicative of future price movements. To achieve this a well-known dynamic topic modelling approach is applied to social media communication to retrieve information about the temporal occurrence of various topics. A Hawkes model is then applied to find interactions between topics and cryptocurrency prices. The results show particular topics tend to precede certain types of price movements, for example the discussion of 'risk and investment vs trading' being indicative of price falls, the discussion of 'substantial price movements' being indicative of volatility, and the discussion of 'fundamental cryptocurrency value' by technical communities being indicative of price rises. The knowledge of topic relationships gained here could be built into a real-time system, providing trading or alerting signals.

Over the past few years, there has been a growth in activity, public knowledge, and awareness of cryptocurrencies and related blockchain technology. As the industry has grown, there has also been an increase in scams looking to steal unsuspecting individuals' cryptocurrency. Many of the scams operate on visually similar but seemingly unconnected websites, advertised by malicious social media accounts, which either attempt an advance-fee scam or operate as phishing websites. This paper analyses public online and blockchain-based data to provide a deeper understanding of these cryptocurrency scams. The clustering technique DBSCAN is applied to the content of scam websites to discover a typology of advance-fee and phishing scams. It is found that the same entities are running multiple instances of similar scams, revealed by their online infrastructure and blockchain activity. The entities also manufacture public blockchain activity to create the appearance that their scams are genuine. Through source and destination of funds analysis, it is observed that victims usually send funds from fiat-accepting exchanges. The entities running these scams cash-out or launder their proceeds using a variety of avenues including exchanges, gambling sites, and mixers.

No abstract