Leverage data on name variants to improve phase two matching about reciter

[michaelbales1]

Based on match between the target and the names in the candidate cluster, apply scores as follows:

	First name	Middle name	Last name
Direct match with target	1	1	1
Abbrev. match with target	0.8	0.9	N/A
No match	0.2	0.3	N/A
Null	N/A	0.5	N/A

[michaelbales1]

Get data from rc_identity_directory

[michaelbales1]

Background

In phase two matching ReCiter determines which, if any, clusters are to be assigned to the target author. The goal of this improvement is to leverage an author’s aliases to improve this selection process.

Operationalization

• Match cluster originator against first, middle, and last name to the target author’s aliases.
• Based on match between the target’s aliases and the names in the candidate cluster, apply scores as in the above table.

[michaelbales1]

#93 must be completed before this one. Also related to #100 and #59.

[michaelbales1]

Hanumantha indicates that he has implemented this code and that Jie has integrated it into the ReCiterAlgorithmJul2015 branch. I will follow up with Jie for clarification.

[michaelbales1]

Hanumantha states this has been implemented in ReCiterClusterer.java at line 594.

[paulalbert1]

Addressed here: https://github.com/wcmc-its/ReCiter/blob/master/src/main/java/reciter/algorithm/evidence/targetauthor/name/RemoveByNameStrategyContext.java