sawnuti: Comparing Sequences with Non-Uniform Time Intervals

In sawnuti: Comparing Sequences with Non-Uniform Time Intervals

Description

The SAWNUTI algorithm performs sequence comparison for finite sequences of discrete events with non-uniform time intervals. Further description of the algorithm can be found in the paper: A. Murph, A. Flynt, B. R. King (2021). Comparing finite sequences of discrete events with non-uniform time intervals, Sequential Analysis, 40(3), 291-313. <doi:10.1080/07474946.2021.1940491>.

Usage

sawnuti(string1, string2, times1, times2, global = T, alpha,
match_function, gap_penalty)

Arguments

string1	character. Character of representations of each element in the first sequence separated by a single white space.
string2	character. Character of representations of each element in the first sequence separated by a single white space.
times1	character. Numerical time values between each value in the first sequence. Should be written as a single character string, where each value is separated by a single white space.
times2	character. Numerical time values between each value in the second sequence. Should be written as a single character string, where each value is separated by a single white space.
global	logical. Logical representing whether a user wishes to perform a global or local alignment. T = global, F = local.
alpha	double. Time interval penalty bias. Weights the influence of time on the alignment calculation.
match_function	R function. Score given for alignment of particular values. Must be able to take in two values from the universe of possible events, and return a numerical score for the alignment of those two elements. Simplest is a constant score returned, can also be implemented as a matrix look-up.
gap_penalty	numerical. This implementation only allows for constant gap penalties.

Details

Zero-one scale of all time values in the entire dataset is assumed before calculation. See Examples for possible formatting of the match_function.

Value

Vector containing 'similarity score', 'alignment', and 'scoring matrix'.

PREPROCESSING ASSUMPTION

Zero-one scale of all time values in the entire dataset is assumed before calculation. Since this preprocessing step must be done with ALL possible time intervals, it cannot be done in this alignment which only takes in two particular observations.

References

A. Murph, A. Flynt, B. R. King (2021). Comparing finite sequences of discrete events with non-uniform time intervals, Sequential Analysis, 40(3), 291-313.

Examples

matchFunction = function(a,b){ifelse(a==b, 1, -1)}

sawnuti(string1="a b c", string2="d b c", times1="1 2 3", 
times2="3 2 1", alpha = 1, match_function = matchFunction, gap_penalty = 1)

sawnuti documentation built on Dec. 13, 2021, 5:07 p.m.