chronofeat is an R package for time-based feature engineering and forecasting. It provides a flexible, formula-based interface for creating temporal features (lags, moving averages, rolling statistics, calendar features) and works with any R model that has a fit/predict interface.
Features
-
Formula-based feature specification: Define features using intuitive syntax like
value ~ p(12) + q(7) + month() - Model-agnostic: Works with lm, glm, xgboost, lightgbm, randomForest, or any custom model
- Recursive multi-step forecasting: Automatically generates features at each forecast step
- Panel data support: Handle multiple time series with proper group boundaries
- C++ acceleration: Fast recursive forecasting via cpp11
About TAFS
TAFS (Time Series Analysis and Forecasting Society) is a non-profit association (“Verein”) in Vienna, Austria. It connects academics, experts, practitioners, and students focused on time-series, forecasting, and decision science. Contributions remain fully open source. Learn more at taf-society.org.
Installation
The development version from GitHub with:
# install.packages("remotes")
remotes::install_github("taf-society/chronofeat")Quick Start
library(chronofeat)
# Load sample data
data(retail)
# Create TimeSeries object
ts_data <- TimeSeries(retail, date = "date", groups = "items", frequency = "month")
# Fit a model with formula-based features
model <- fit(
value ~ p(12) + q(7, 12) + month(),
data = ts_data,
model = lm
)
# Generate forecasts
forecasts <- forecast(model, h = 6)Formula Syntax
| Syntax | Description |
|---|---|
p(k) |
k lags of target |
p(1, 7, 12) |
Specific lags |
q(w1, w2) |
Moving averages |
dow() |
Day of week |
month() |
Month |
rollsum(w) |
Rolling sum |
rollsd(w) |
Rolling std dev |
trend(d) |
Polynomial trend |
lag(var, k) |
Lag of exogenous |
Documentation
Visit the package website for: