D-Net
/
dnet-dedup
5
1
Fork 0
Code Issues Pull Requests 1 Releases Wiki Activity

minor change in the author match which now can compute count and percentage

This commit is contained in:
Michele De Bonis 2023-04-04 17:10:37 +02:00
parent b4b6a61576
commit b5584f084a
1 changed files with 8 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
dnet-pace-core/src/main/java/eu/dnetlib/pace/tree/AuthorsMatch.java
View File

@ -26,6 +26,7 @@ public class AuthorsMatch extends AbstractComparator {
private double FULLNAME_THRESHOLD;
private String MODE; //full or surname
private int SIZE_THRESHOLD;
private String TYPE; //count or percentage
private int common;
public AuthorsMatch(Map<String, String> params){
@ -37,6 +38,7 @@ public class AuthorsMatch extends AbstractComparator {
NAME_THRESHOLD = Double.parseDouble(params.getOrDefault("name_th", "0.95"));
FULLNAME_THRESHOLD = Double.parseDouble(params.getOrDefault("fullname_th", "0.9"));
SIZE_THRESHOLD = Integer.parseInt(params.getOrDefault("size_th", "20"));
TYPE = params.getOrDefault("type", "percentage");
common = 0;
}
@ -123,8 +125,13 @@ public class AuthorsMatch extends AbstractComparator {
//normalization factor to compute the score
int normFactor = aList.size() == bList.size() ? aList.size() : (aList.size() + bList.size() - common);
if(TYPE.equals("percentage")) {
return (double) common / normFactor;
}
else {
return (double) common;
}
}
public boolean compareSurname(Person p1, Person p2) {
return ssalgo.score(normalization(p1.getNormalisedSurname()), normalization(p2.getNormalisedSurname())) > SURNAME_THRESHOLD;