We investigate the predictability of Bitcoin, Ripple and Litecoin markets using Twitter data and machine learning models from National Neural Network (NN) and Recurrent Neural Network (RF) sources. Our analyses indicate that Twitter data alone can better anticipate market movements than either of the latter two approaches.
Cryptocurrencies have recently gained increasing attention due to their elevated volatility and low barrier for entry. Research into their quantile connectedness, contagion, and tail spillover with traditional financial markets (Umar and Gubareva 2020) continues to expand rapidly.
Market Conditions
Cryptocurrency prices are highly volatile and exhibit poor market efficiency when measured against Fama’s efficient frontier hypothesis (Paradiso et al. 2021). Due to its unique characteristics, cryptocurrency markets cannot be directly compared to conventional ones; their response to positive and negative shocks being far greater than with conventional financial assets.
This paper employs the cross-quantilogram method to explore quantile dependence between five leading cryptocurrencies and three volatility indices of equity, oil and gold markets. Furthermore, time-varying network connectedness and relative tail dependency based on pairwise correlations is explored as well as pairwise network connectivity in terms of magnitude and direction.
Our results reveal that systemic risk transmission and market contagion are greatly influenced by market conditions. For instance, the COVID-19 pandemic caused high levels of uncertainty and led to bear markets, weakening systemic connectedness. As depicted in Figure 12, four net transmitters such as ETH, LTC, BCH, and XRP transmit innovations to both the OVX and GVZ volatility indices via panel A of Figure 12.
Predictions
Cryptocurrency prices are determined by a range of factors, including economic and interest rate dynamics and political developments. Forecasts can help traders and investors make informed decisions when to buy or sell. It’s important to remember that cryptocurrency is a highly speculative market with highly volatile prices; before making any investments it’s essential to do your own research to assess your level of risk tolerance before taking the plunge.
One can employ various strategies to predict cryptocurrency prices, from building predictive models based on statistics to training algorithms to find patterns. Combining different techniques will improve prediction accuracy while decreasing risks associated with relying solely on one technique; nonetheless, no method is foolproof as unexpected events may have an unexpected impact on price volatility and crypto prices – this makes diversification essential in managing risks effectively.
Conclusions
Cryptocurrency predictions must be treated with extreme skepticism, especially those made by those claiming they know how prices will behave. Due to its infancy and numerous unproven factors that could impact market behavior, potential investors should always perform their own research before taking an investment decision.
Moving averages and Fibonacci retracement ratios are widely-used trading indicators that help predict cryptocurrency price trends. These mathematical interpretations of historical price activity and trading volume provide insights into which trends to pursue.
Gurdgiev and O’Loughlin (2020) studied the relationship between cryptocurrency price dynamics and fear (VIX index), uncertainty (U.S. equity market uncertainty index), investor sentiment towards cryptocurrency (measured from Bitcoin forums), and investor behavior to conclude that investor sentiment greatly influences price dynamics of cryptocurrencies similar to traditional markets. As with traditional markets, sentiment can significantly impact cryptocurrency price dynamics as well.
References
Amid the widespread interest in cryptocurrencies, discussions of risk transmission, contagion and tail spillover have increased substantially (Fendel and Neumann 2021; Umar and Gubareva 2022). This research adds to existing literature by investigating time-varying cross-quantilogram network connectedness of leading cryptocurrencies with equity, oil and gold volatility indices in traditional markets. We find that BTC returns positively predict ETH, LTC and XRP returns when controlling for market uncertainty through equity oil gold volatility indices; predictability is strongest when reflecting normal crypto market conditions but weakens further when bearish or bullish conditions occur.
Ripple is a cryptocurrency designed to address the challenge of expensive and slow international payments by enabling banks to send small amounts quickly and affordably across borders. Based on blockchain technology, it has already been tested with several large financial institutions; its price fluctuates often as with all cryptocurrencies.
