Deprecated in bibliometrix 6.0. This function has been renamed to
applyReferenceMatching. Please use applyReferenceMatching() instead.
applyCitationMatching(
M,
threshold = 0.9,
method = "jw",
min_chars = 20,
max_block_size = 100,
use_iso4 = TRUE,
use_doi = TRUE,
use_exact = TRUE,
fuzzy = TRUE,
use_postproc = TRUE,
title_guard = FALSE
)Same as applyReferenceMatching.
A bibliometrix data frame, typically created by convert2df.
Must contain the columns:
SR: Short reference identifier for each document
CR: Cited references field (citations separated by semicolons)
DB: (Optional) Database source identifier for format detection
Numeric value between 0 and 1 indicating the similarity threshold
for matching citations. Default is 0.90. See normalize_citations
for details on selecting appropriate thresholds.
String distance method to use for fuzzy matching. Options include:
"jw" (default): Jaro-Winkler distance, optimized for bibliographic strings
"lv": Levenshtein distance
"osa": Optimal String Alignment distance
"lcs": Longest Common Subsequence distance
Other methods supported by stringdistmatrix
Minimum characters for valid citations (default: 20).
Integer. Blocks with at least this many unique normalized
strings skip within-block fuzzy matching and fall back to exact matching only
(default: 100). Passed to normalize_citations.
Logical. Enable ISO 4 / LTWA journal normalization (default TRUE).
Logical. Enable DOI-based exact matching (default TRUE).
Logical. Enable exact normalized-string and punctuation-invariant
matching (default TRUE).
Logical. Enable within-block (WoS deterministic and Scopus fuzzy)
matching (default TRUE).
Logical. Enable Phase 4.5 metadata-based post-processing
merge (default TRUE). These five logical switches are intended primarily
for ablation analyses and are passed through to normalize_citations.
Logical. Enable the optional Phase 4.6 series part-marker
cluster purification that splits over-merged multi-part works (default
FALSE). Passed through to normalize_citations.