This function returns a trimmed COVLMC from which cached data have been removed.
Usage
# S3 method for covlmc
trim(ct, keep_model = FALSE, ...)Arguments
- ct
a context tree.
- keep_model
specifies whether to keep the internal models (or not)
- ...
additional arguments for the trim function.
Details
Called with keep_model set to FALSE (default case), the trimming is maximal and reduces
further usability of the model. In particular loglikelihood.covlmc() cannot be used
for new data, contexts.covlmc() do not support model extraction, and
simulate.covlmc(), metrics.covlmc() and prune.covlmc() cannot be used at all.
Called with keep_model set to TRUE, the trimming process is less complete. In
particular internal models are simplified using butcher::butcher() and some
additional minor reductions. This saves less memory but enables the use of
loglikelihood.covlmc() for new data as
well as the use of simulate.covlmc().
Examples
pc <- powerconsumption[powerconsumption$week %in% 5:7, ]
dts <- cut(pc$active_power, breaks = c(0, quantile(pc$active_power, probs = c(0.5, 1))))
dts_cov <- data.frame(day_night = (pc$hour >= 7 & pc$hour <= 17))
m_cov <- covlmc(dts, dts_cov, min_size = 10, keep_data = TRUE)
print(object.size(m_cov), units = "Mb")
#> 26.9 Mb
t_m_cov_model <- trim(m_cov, keep_model = TRUE)
print(object.size(t_m_cov_model), units = "Mb")
#> 8.3 Mb
t_m_cov <- trim(m_cov)
print(object.size(t_m_cov), units = "Mb")
#> 0.1 Mb