This function prunes a vlmc with covariates. This model must have been
estimated with keep_data=TRUE to enable the pruning.
Usage
# S3 method for covlmc
prune(vlmc, alpha = 0.05, cutoff = NULL, ...)Arguments
- vlmc
a fitted VLMC model with covariates.
- alpha
number in (0,1) (default: 0.05) cutoff value in quantile scale for pruning.
- cutoff
not supported by the vlmc with covariates.
- ...
additional arguments for the prune function.
Details
Post pruning a VLMC with covariates is not as straightforward as the same
procedure applied to vlmc() (see cutoff.vlmc() and prune.vlmc()). For
efficiency reasons, covlmc() estimates only the logistic models that are
considered useful for a given set construction parameters. With a more
aggressive pruning threshold, some contexts become leaves of the context tree
and new logistic models must be estimated. Thus the pruning opportunities
given by cutoff.covlmc() are only a subset of interesting cut offs for a
given covlmc.
Nevertheless, covlmc share with vlmc() the principle that post pruning a
covlmc should give the same model as buidling directly the covlmc, provided
that the post pruning alpha is smaller than the alpha used to build the
initial model.
Examples
pc <- powerconsumption[powerconsumption$week == 5, ]
dts <- cut(pc$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 = 5, keep_data = TRUE)
draw(m_cov)
#> *
#> +-- (0,1.34]
#> | +-- (0,1.34]
#> | | +-- (0,1.34] (0.02329 [ -2.813 2.178 -4.449 2.418 ])
#> | | '-- (1.34,7.54] (0.01834 [ -19.57 18.39 ])
#> | '-- (1.34,7.54] (0.6763 [ -1.856 ])
#> '-- (1.34,7.54] (0.8175 [ 2.535 ])
m_cov_cuts <- cutoff(m_cov)
p_cov <- prune(m_cov, m_cov_cuts[1])
draw(p_cov)
#> *
#> +-- (0,1.34]
#> | +-- (0,1.34] (0.02001 [ -2.759 2.785 -2.525 ])
#> | '-- (1.34,7.54] (0.6763 [ -1.856 ])
#> '-- (1.34,7.54] (0.8175 [ 2.535 ])