backtrack {outlierMBC}R Documentation

Move backwards from the minimum to a more conservative solution.

Description

Given a vector of dissimilarity values, each corresponding to a different number of outliers, this function first finds the index and value of the minimum dissimilarity, then moves backwards from right to left to a reasonable solution with a lower index (i.e. lower number of outliers). Limits are placed on the maximum increase in dissimilarity from a single step (max_step_rise) and from all steps (max_total_rise), where both are defined in proportion to the minimum dissimilarity value.

Usage

backtrack(x, max_total_rise = 0.1, max_step_rise = 0.05)

Arguments

x

Vector of dissimilarity values corresponding to consecutive and increasing numbers of outliers.

max_total_rise

Upper limit for the cumulative increase, as a proportion of the global minimum dissimilarity, from all backward steps.

max_step_rise

Upper limit for the increase, as a proportion of the global minimum dissimilarity, from each backward step.

Value

backtrack returns a list with two elements, minimum and backtrack:

minimum is a list with the following elements:
ind

Index of the minimum solution.

val

Value of the minimum solution.

backtrack is a list with the following elements:
ind

Index of the backtrack solution.

val

Value of the backtrack solution.

Examples

ombc_gmm_k3n1000o10 <-
  ombc_gmm(gmm_k3n1000o10[, 1:2], comp_num = 3, max_out = 20)

backtrack(ombc_gmm_k3n1000o10$distrib_diff_vec)

[Package outlierMBC version 0.0.1 Index]