Cryptocurrency data firm Nomics is applying artificial intelligence to the often chaotic world of cryptocurrency trading.
In an announcement on April 23, Nomics introduced an AI system that makes seven-day cryptocurrency price predictions. The firm claims this makes it the first to share crypto asset price predictions generated by a machine learning (ML)-based AI system.
Clay Collins, the firm’s CEO, told Cointelegraph that the predictions are meant for retail investors looking for ML predictions, and should not be taken as gospel:
“These predictions are not investment advice. Investments and trading in crypto assets involve substantial risk of loss and is not suitable for every investor. These predictions should be used as one of several indicators to inform opinions about future price movements.”
Collins believes that this kind of AI can — and will — do more than just guess the future price of assets:
“Machine learning will increasingly be layered on top of crypto market data — and financial data in general — to spot fake volume, wash trading, and to make informed predictions about the future. Today, we're taking our first step towards using machine learning to elucidate crypto data trends.”
Nomics’ price prediction system is based on a long short-term memory (LSTM) ML model. ML systems are not programmed in the way traditional software is, instead they learn by examining large quantities of data. In this case, the AI was trained using the cryptocurrency price data accumulated by Nomics during its activity.
Error data will be published alongside the predictions
The AI feeds the price predictions back to Nomics so it can be displayed on the website. The firm’s chief technology officer Nick Gauthier promised that the firm will be open when it comes to the AI’s error rate:
“LSTMs are relatively new in the machine learning space, and financial data is notoriously difficult to predict, but we were able to get fairly reasonable predictions, and we are fully transparent about their historical accuracy.”
Nomics' website shows the predicted seven-day percentage change from the current price, the seven-day predicted price, and the mean 30-day error. Collins noted:
“We know that predictions can be wrong. [...] But we felt that if we could find a transparent way to do this, then it could be a beneficial signal — along with your own research — for making sense of the Cryptoverse and related price movements.”
As a recent Cointelegraph analysis explains, companies are increasingly working to improve data transparency and accountability — key components for onboarding institutional investors to cryptocurrency markets.