Improve method for identifying target author

[paulalbert1]

Every paper - including those from the gold standard and those where the ReCiter algorithm opts not to make a match - should have one and only one target author.

All matching should be done in a case-insensitive way.

The output of this judgment should be:

• done at an early state in the algorithm
• stored in the Analysis table (which hasn't been created yet)
• used for analysis
• used as a field to be output into the feature generator API.

Some of this work is being done by getCorrectAuthor.

Order of operations

• Retrieve records from PubMed and according to the range of retrieval strategies
• Get complimentary data from Scopus
• Identify targetAuthor

Logic

Attempt email match

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result (e.g., 28265069 for OS Andersen), assign remaining authors as false, go to next.

Attempt strict last name, strict middle name, and strict first name match.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt strict last name, middle initial, and strict first name match.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt strict last name and strict first name match.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt strict last name and partial first name match, in which article is substring of identity.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt strict last name and partial first name match, in which identity is substring of article.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt strict last name and first initial match.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt strict last name and middle initial to first initial, and first initial to middle initial match.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt partial last name and first initial match.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt strict last name match.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt strict first name and first initial of last name match.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt strict first name match.

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

Attempt full last name match from article to partial last name from identity. (e.g., Somersan-Karakaya)

• if 0 results, go to next.
• if 1 result, stop.
• if > 1 result, assign remaining authors as false, go to next.

If one of the names in Article is for a “Collective Author” (examples - this is for rbdevere), map to that.

• if 0 results, assign all authors as false.
• if 1 result, stop. Assign that author as true.
• if > 1 result, assign ALL authors as false, go to next.

Assign all authors as false.

[sarbajitdutta]

The last step to check for collective author is ON HOLD since we need to bring in additional data from pubmed namely Contributor list

[paulalbert1]

@sarbajitdutta - For ses9022, I got this:

19:37:49.914 [http-nio-5000-exec-4] INFO  r.a.e.t.TargetAuthorSelection - There was no target author found for 29117776
19:37:49.947 [http-nio-5000-exec-4] INFO  r.a.e.t.TargetAuthorSelection - There was no target author found for 27398323
19:37:49.950 [http-nio-5000-exec-4] INFO  r.a.e.t.TargetAuthorSelection - There was no target author found for 28352671
19:37:49.967 [http-nio-5000-exec-4] INFO  r.a.e.t.TargetAuthorSelection - There was no target author found for 28287838

---

---

For ajg9004, I got this:

07:50:05.603 [http-nio-5000-exec-9] INFO  r.a.e.t.TargetAuthorSelection - 3 authors were marked as target author for article 29603699

This seems doable...

        "rank": 1,
          "lastName": "Gupta",
          "firstName": "Ajay K",
          "initials": "A",
          "affiliations": null,
          "targetAuthor": true
        },
        {
          "rank": 4,
          "lastName": "Gupta",
          "firstName": "Amit K",
          "initials": "A",
          "affiliations": null,
          "targetAuthor": true
        },
        {
          "rank": 6,
          "lastName": "Gupta",
          "firstName": "Deepika",
          "initials": "D",
          "affiliations": null,
          "targetAuthor": true
        }

Is the above logic wrong??

[paulalbert1]

I've noticed that email match is not very effective. We might consider commenting it out - or moving down to position 6 in the rubric.

[paulalbert1]

No target author for aae2001

{
      "rank": 1,
      "lastName": "El Menyar",
      "firstName": "Ayman A",
      "initials": "A",
      "targetAuthor": false
    },


"primaryName": {
  "firstInitial": "A",
  "firstName": "Ayman",
  "lastName": "El-menyar",
  "middleInitial": "A",
  "middleName": "A."
},        

PMID = 15590364

[sarbajitdutta]

Add deAccent to the AuthorSanitizationUtil.

[paulalbert1]

Here's a tricky case
amr2011 - https://www.ncbi.nlm.nih.gov/pubmed/27922594

Should match to

<LastName>Rajadhyaksha</LastName>
<ForeName>Anjali M</ForeName>

[paulalbert1]

Fixed.