localScore (version 2.0.1)

localScoreC: Local score

Description

Calculates the local score for a sequence of scores, the sub-optimal segments, and the associated record times. The local score is the maximal sum of values contained in a segment among all possible segments of the sequence. In other word, it generalizes a sliding window approach, considering of all possible windows size.

Usage

localScoreC(v, suppressWarnings = FALSE)
localScoreC_double(v, suppressWarnings = FALSE)
localScoreC_int(v, suppressWarnings = FALSE)

Value

A list containing:

localScore

the local score value and the begin and end index of the segment realizing this optimal score;

suboptimalSegmentScores

An array containing sub-optimal local scores, that is all the local maxima of the Lindley rocess (non negative excursion) and their begin and end index;

RecordTime

The record times of the Lindley process as defined in Karlin and Dembo (1990).

Arguments

v

a sequence of numerical values as vector (integer or double).

suppressWarnings

(optional) if warnings should not be displayed

Details

The localScoreC function is implemented in a templated C function. Be aware that the type of the output (integer or double) depends on the type of the input. The function localScoreC_double localScoreC_int explicitly use the corresponding type (with an eventual conversion in case of integer). Warning: in R, typeof(c(1,3,4,10)) == "double". You can set a type of a vector with mode() or as.integer() functions for example.
localScoreC_int is just a call to as.integer() before calling localScoreC. localScore_double is just a call to localScoreC, and as such is deprecated.

See Also

lindley

Examples

localScoreC(c(1.2,-2.1,3.5,1.7,-1.1,2.3))
# one segment realizing the local score value
seq.OneSegment <- c(1,-2,3,1,-1,2)
localScoreC(seq.OneSegment) 
seq.TwoSegments <- c(1,-2,3,1,2,-2,-2,-1,1,-2,3,1,2,-1,-2,-2,-1,1)
# two segments realizing the local score value
localScoreC(seq.TwoSegments) 
# only the first realization
localScoreC(seq.TwoSegments)$localScore 
# all the realization of the local together with the suboptimal ones
localScoreC(seq.TwoSegments)$suboptimalSegmentScores 
# for small sequences, you can also use lindley() function to check if 
# several segments achieve the local Score
lindley(seq.TwoSegments) 
plot(1:length(seq.TwoSegments),lindley(seq.TwoSegments),type='b')
seq.TwoSegments.InSameExcursion <- c(1,-2,3,2,-1,0,1,-2,-2,-4,1)
localScoreC(seq.TwoSegments.InSameExcursion)
# lindley() shows two realizations in the same excursion (no 0 value between the two LS values)
lindley(seq.TwoSegments.InSameExcursion) 
# same beginning index but two possible ending indexes
# only one excursion realizes the local score even in there is two possible lengths of segment
localScoreC(seq.TwoSegments.InSameExcursion)$suboptimalSegmentScores 
plot(1:length(seq.TwoSegments.InSameExcursion),lindley(seq.TwoSegments.InSameExcursion),type='b')
# Technical note about type correspondance
seq.OneSegment <- c(1,-2,3,1,-1,2)
seq.OneSegmentI <- as.integer(seq.OneSegment)
typeof(seq.OneSegment)  # "double" (beware)
typeof(seq.OneSegmentI) # "integer"
LS1 <- localScoreC(seq.OneSegment, suppressWarnings = TRUE)
LS1I <- localScoreC(seq.OneSegmentI, suppressWarnings = TRUE)
typeof(LS1$localScore)  # "double"
typeof(LS1I$localScore) # "integer"
typeof(LS1$suboptimalSegmentScores$value)  # "double"
typeof(LS1I$suboptimalSegmentScores$value) # "integer"
# Force to use integer values (trunk values if necessary)
seq2 <- seq.OneSegment + 0.5
localScoreC(seq2, suppressWarnings = TRUE)
localScoreC_int(seq.OneSegment, suppressWarnings = TRUE)