Skip to content

predict the future of time series data

Notifications You must be signed in to change notification settings

Vraid-Systems/ml_timeseries

Repository files navigation

ml_timeseries

Use machine learning to predict the next values in the timeseries. With enough of the right input, future patterns can be predicted.

Packaging

docker build --tag docker.pkg.github.com/vraid-systems/ml_timeseries/main:`cat DOCKER_IMAGE_VERSION` .

Usage

Create predictions from Quadency cryptocurrency average price API for listed cryptocurrencies:

docker run --env CRYPTO=BTC/USD,ETH/USD --env MODEL=arima --volume /local/path/to/prediction.json:/tmp/ml_timeseries/prediction.json docker.pkg.github.com/vraid-systems/ml_timeseries/main:3.0.0

Create prediction from Yahoo Finance stock ticker API for particular stock:

docker run --env MODEL=lstm --env STOCK=JNJ /local/path/to/prediction.json:/tmp/ml_timeseries/prediction.json docker.pkg.github.com/vraid-systems/ml_timeseries/main:2.0.0

Note that MODEL can be arima or lstm.