complete_cases {onlineforecast}

Find complete cases in forecast matrices

Description

Returns a logical vector indicating the time points which

Usage

complete_cases(object, kseq = NA)

## S3 method for class 'list'
complete_cases(object, kseq = NA)

## S3 method for class 'data.frame'
complete_cases(object, kseq = NA)

Arguments

object

A data.frame (with columns named 'kxx') or a list of data.frames.

kseq

integer vector: If given then only these horizons are processed.

Details

Given a forecast matrix the forecasts are lagged "+k" steps to align them and then 'complete.cases()' is run on that .

Gieven a list of forecast matrices the points where all are complete (also all horizons) are complete are TRUE.

Value

A logical vector specifying if there is no missing values across all horizonsd.

Author(s)

Peder Bacher

Examples

# Take a small data set
D <- subset(Dbuilding, 1:20, kseq=1:5)
# Check the forecast matrix of ambient temperature
D$Ta
# Which are complete over all horizons? The first are not since not all horizons
# have a value there (after lagging)
complete_cases(D$Ta)
# Same goes if given as a list
complete_cases(D["Ta"])
# and if more than one is given
complete_cases(D[c("Ta","I")])

# Set some NA of some horizon
D$I$k3[8:9] <- NA
# Now they are recognized as not complete
complete_cases(D[c("Ta","I")])

# If we deal with residuals, which are observations and there for have column names "hxx"
Resid <- residuals(D$Ta, D$Taobs)
names(Resid)
# With columns with "h" instead of "k" no lagging occurs in complete_cases
complete_cases(Resid)
#
Resid2 <- Resid
Resid$h3[8:9] <- NA
complete_cases(list(Resid,Resid2))