FreqAI: generalizing adaptive modeling for chaotic time-series market forecasts

FreqAI evolved from a desire to rapidly test and compare a range of adaptive time-series forecasting methods on live chaotic data. Cryptocurrency markets provide a unique chaotic data source as they are operational 24/7 and the data is freely available via a variety of open-sourced exchange APIs. Luckily, an existing open-source software, Freqtrade, had already matured under a range of talented developers to support robust data collection/storage, as well as robust live environmental interactions for standard algorithmic trading. Freqtrade also provides a set of data analysis/visualization tools for the evaluation of historic performance as well as live environmental feedback. FreqAI builds on top of Freqtrade to include a user-friendly well tested interface for integrating a wide variety of external machine learning libraries for adaptive time-series forecasting. Beyond enabling the integration of existing libraries, FreqAI hosts a range of custom algorithms and methodologies aimed at improving computational and predictive performances. Thus, FreqAI contains a range of unique features that can be easily tested in combination with all the existing Python-accessible machine learning libraries to generate novel research on live and historical data.

The target user of FreqAI is someone who would like to rapidly test and compare hypotheses related to adaptive modeling in live and historic chaotic market environments. These users typically aim to compare and contrast across a broad range of open-source machine learning libraries, such as PyTorch, CatBoost, XGBoost, LightGBM, Tensorflow, Scikit-learn, etc. Users do not need to be fluent in in all aforementioned libraries, since FreqAI provides starting templates for each, as well as example configuration files and exhaustive documentation.

Development and latest releases of FreqAI are all accessible in the github repository online. The sources archived here (version JOSS_v1) were reviewed and accepted by the Journal of Open Source Software (JOSS). Find details of the review process here.