sousrir_ssdtw: Return score of how likely a query occurs in a reference,...

In parledoct/sousrir: Search On Untranscribed Speech Rapidly In R

Description

Return score of how likely a query occurs in a reference, given a starting index

Usage

sousrir_ssdtw(
  query_name,
  ref_name,
  query_feats,
  ref_feats,
  top_match_start,
  min_match_ratio = 0.5,
  max_match_ratio = 2,
  distance_func = dist_scipy_stdeuc,
  distnorm_func = norm_rf2014,
  return_dtwalign = FALSE
)

Arguments

query_name	Name of query (for error reporting in case function fails)
ref_name	Name of reference (for error reporting in case function fails)
query_feats	Query feature matrix (of shape M rows and F columns)
ref_feats	Reference feature matrix (of shape N rows and F columns)
top_match_start	Top match location returned by sousrir_1nndtw
min_match_ratio	Minimum match length as ratio of query (default: 0.5 = half the query size)
max_match_ratio	Maximum match length as ratio of query (default: 2.0 = twice the query size)
distance_func	Function to compute distances between query and reference (default: dist_stdeuc)
distnorm_func	Function to normalize computed distances (default: norm_rf2014)
return_dtwalign	Whether or not to return alignment object (i.e. for plotting alignment)

Value

A numeric score of how likely the query occurs in the reference (or NA if no acceptable alignment was found given the parameters)

parledoct/sousrir documentation built on Dec. 22, 2021, 6:39 a.m.