Predict the baseline intensity trend at one or more future time points

Decomposes the estimated piecewise-constant log-baseline (est_time) into a smooth trend component (via LOESS) and a seasonal/residual component, following the same approach used in the application plot script. The fitted trend is then extrapolated to target_times using predict.loess() so that the baseline used for out-of-sample scoring reflects the long-run level of activity rather than any arbitrary fixed value.

Usage

predict_baseline_trend(model, target_times, loess_span = 0.75)

Arguments

model

A redeem_result object with non-null est_time and time_changepoints fields.

target_times

Numeric vector: the times at which to predict the trend (typically the unique timestamps of the test events).

loess_span

Numeric; the span argument passed to stats::loess. Larger values give a smoother (more conservative) trend extrapolation. Defaults to 0.75.

Value

A numeric vector of predicted log-baseline (trend component) values, one per element of target_times. Falls back to mean(est_time) for each time point if there are fewer than 3 observations or if the LOESS fit fails.

Details

The decomposition mirrors the plot code in the application:

1. Build a data frame of (time, est_time) using the changepoints stored in model$time_changepoints. The first interval [0, changepoint_1) is given time = 0; each subsequent interval gets the corresponding changepoint value.

2. Fit LOESS on the log-scale est_time values.

3. Predict at each target_times; predictions are clamped to the range of the observed est_time to avoid wild extrapolation.