This brand new Python library GreyKite is released by Linkedin . It is used for time series forecasting. This library makes the life of data scientists easier. This library provides automation with the help of the Silverkite algorithm . LinkedIn created GrekKite to help its group settle on viable choices dependent on the time-series forecasting models. This also helps to interpret the outputs.
Algorithms supported by Gryekite library
- Silverkite (Greykite’s flagship algorithm)
- Facebook Prophet
Throughout the long term, LinkedIn has been utilizing the Greykite library to give an adequate foundation to deal with top traffic, set business targets, and advance spending choices.
Features of GreyKite
Fast training and scoring
- Works with intelligent prototyping, framework search, and benchmarking. Matrix search is valuable for model choice and self-loader estimating of different measurements.
- Gives time arrangement regressors to catch pattern, irregularity, occasions, changepoints, and autoregression, and allows you to add your own.
- Fits the conjecture utilizing an AI model.
- Gives incredible plotting apparatuses to investigate irregularity, connections, changepoints, and so forth
- Gives model formats (default boundaries) that function admirably dependent on information attributes and gauge prerequisites (for example every day long haul conjecture).
- Produces interpretable yield, with model rundown to analyze individual regressors, and segment plots to outwardly review the consolidated impact of related regressors
- Uncovered numerous estimate calculations in a similar interface, making it simple to attempt calculations from various libraries and think about outcomes.
- A similar pipeline gives preprocessing, cross-approval, backtest, estimate, and assessment with any calculation.