The limitations of artificial intelligence (AI) capabilities are still unclear. Recently, new evidence has emerged through research conducted by researchers at the Indraprastha Institute of Information Technology (IIIT) Delhi, who merged artificial intelligence and traditional finance to predict cryptocurrency prices.
Recently, a prestigious academic journal has accepted this research. In fact, the Elsevier Information Science corroborates the effectiveness of the method developed by the Delhi researchers to predict cryptocurrency prices. This study was made possible by Shalini Sharma, a PhD student, and Dr Angshul Majumdar, her supervisor.
Generally, there are two main pricing methods in financial markets. One is based on traditional methods from the 1970s. This is the Baum-Welch probability method, which allows for price prediction in financial markets, in addition to uncertainty about the prediction.
However, this approach has its limitations in relation to cryptocurrencies, since there is no data available on the factors that cause price changes. This is where the modern perspective brings additional value.
Based on Artificial Intelligence, this method is also known as deep learning. However, the latter does not require any essential information or knowledge. Therefore, it does not yet cover the uncertainty variable linked to cryptocurrencies.
The finding prior to this study is that there were no custom methods to incorporate uncertainty into cryptocurrency predictions.
Shalini Sharma and Angshul Majumdar have developed a novel strategy that addresses the problem of prediction uncertainty linked to cryptocurrencies. With the information available, this is the best possible in terms of accuracy. Moreover, this method outperforms all existing modern techniques.
For a better understanding of the method suggested by the Delhi researchers, it is essential to clarify the concept of CVI. The CVI is simply the crypto volatility index. For this, the variability of a crypto over time is considered. For instance, a stablecoin like DAI or USDC would have a lower index compared to cryptos like Dogecoin or Shiba Inu.
As a result, the uncertainty estimates achieved by the two researchers are related to the historical CVI values of the available cryptos. Therefore, the conclusions obtained through this new method are indeed understandable.
These two scientists are proposing a real revolution. In fact, the prediction techniques available prior to these studies were mainly based on traditional strategies for the analysis of financial markets. However, cryptos, like the blockchain from which they come, are challenging established paradigms. This is another advance in the digital revolution that Artificial Intelligence is bringing to us.