CurPay AI Models
These example regression models are used in CurPay's Automated Volatility Protection (AVP)® software for smart trades in their Trader app and smart conversions to FIAT in their Merchant app.
They are the 1st stage of a 4 stage AI process that are trained using ML.net.
In this stage the next rate (label) is determined by passing in the following indicators (features): EMA10, EMA12, EMA17, EMA21, EMA100, EMA200, EMA26, MACD, Signal, RSI
For Help in understanding these indicators please refer to https://www.investopedia.com
The CurPay system trains each model using either LightGbm or FastForest regression trainers. During every train it is determined which is the most effective trainer based on the following,
- Past performance
- Rsquared
- Mean Cross Validation Rsquared
- Mean Absolute Error
- Mean Squared Error
- Root Mean Squared Error
- Loss Function
Each file is named based on the Exchange, CurrencyPair and Long Term Short Term
CurPay_{ExchangeID}{BaseCurrencyCode}{ToCurrencyCode}_{LT_ST}.mlnet For example Kraken BTC/USD Long Term outlook would be CurPay_3_BTC_USD_LT.mlnet
For help on consuming these models please refer to Microsoft's ML.net (https://dotnet.microsoft.com/en-us/apps/machinelearning-ai/ml-dotnet)