R/forecasting-gam.R

In ha0ye/MATSS-forecasting: Analysis and Synthesis of Forecasting Ecological Time Series

#' @name gam_ts
#' @title Make forecasts using a Generalized Additive Model
#'
#' @description Fit a Generalized Additive Model (GAM) using 
#'   \code{\link[mgcv]{gam}} with a spline over time/years
#' 
#' @inheritParams forecast_iterated
#' @inheritParams forecast::forecast
#'
#' @return a data.frame of the mean forecasts, the observed values, and the
#'   lower and upper CI levels (if an error occurs, then just NA values)
#' 
#' @export
#'
gam_ts <- function(timeseries, num_ahead = 5, level = 95)
{
    f <- function(training, observed, order, level)
    {
        # make forecasts
        t <- seq_len(length(training))
        gam_model <- mgcv::gam(training ~ s(t)) # gam model with spline over time
        
        # predict function requires a new list of predictor variables as newdata
        t_observed <- length(training) + seq_len(length(observed))
        forecasts <- mgcv::predict.gam(gam_model,
                                       newdata = list(t = t_observed),
                                       se.fit = TRUE)
        
        # return
        return_forecasts(observed, forecasts, level)
    }
    
    forecast_iterated(fun = f, timeseries = timeseries, num_ahead = num_ahead,
                   order = order, level = level)
}