Can we use XGboost to do predition in time series data ? Could anyone explain the logic of mathematics behind that ?
1 Answer
$\begingroup$
$\endgroup$
Yes, I recommend reviewing this package, its documentation and citations:
https://github.com/ellisp/forecastxgb-r-package
Keep in mind, XGB does not typically outperform current time-series methods such as exponential smoothing or ARIMA.