Create journalCategoryScore

[paulalbert1]

Faculty in certain departments are far more likely to publish in journals associated with one or two specialties compared to another. For example, Weill Cornell faculty in Orthopaedic Surgery are 200 times more likely to author papers in journals assigned to the "Orthopedics" subfield/category (as assigned by ScienceMetrix) compared to random chance. We can take advantage of this during scoring.

While there are tens of thousands of combinations between department and subfield, only 100 departments-subfield combinations represent 57% of all known WCM publications. For full validity, we would need to evaluate this approach in other institutions.

Preprocessing

As a one-time thing, we need to create two new static tables into DynamoDB.

• SciMetrix-Revised-2018-08-29.txt
• ScienceMetrixDepartmentCategory.txt
• SciMetrixDepartmentCategory.txt (no PK)

Scoring

1. Get all ISSNs from article.

2. Get identity.organizationalUnits for targetAuthor. Include all types including programs.

3. Look up any department synonyms in application.properties. For example:

departmentSynonym: Healthcare Policy & Research|Public Health|Health Care Policy & Research, Radiation Oncology|Radiation for Oncology, Otolaryngology - Head and Neck Surgery|Otolaryngology|Otorhinolaryngology

For example, support a user has known department of "Public Health", we would need to look in ScienceMetrixDepartmentCategory for departments matching "Healthcare Policy & Research" and "Health Care Policy & Research."

4. Does ISSN exist in the ScienceMetrix table? Check the following possible matches:

• Article.ISSNLinking = ScienceMetrix.ISSN
• Elseif Article.ISSNLinking = ScienceMetrix.eISSN
• Elseif Article.ISSNPrint = ScienceMetrix.ISSN
• Elseif Article.ISSNPrint = ScienceMetrix.eISSN
• Elseif Article.ISSNElectronic = ScienceMetrix.ISSN
• Elseif Article.ISSNElectronic = ScienceMetrix.eISSN

**

• If yes, get journalSubfieldLabel and journalSubfieldID. Go to 5.
• If no, output the following:

journalSubfieldScienceMetrixLabel: null
journalSubfieldScienceMetrixID: null
journalSubfieldDepartment: Anesthesiology
journalSubfieldScore: null

5. Does any organizationalUnit exist for the subfield in question?

• If yes, go to 6.
• If no, output

journalSubfieldScienceMetrixLabel: Biochemistry & Molecular Biology
journalSubfieldScienceMetrixID: 86
journalSubfieldDepartment: Anesthesiology
journalSubfieldScore: -1 /* stored in application.properties */

6. There's a match on subfield and department.

Multiply logOddsRatio by some factor as stored in application.properties:

journalSubfieldFactorScore: 1

Output the following:

journalSubfieldScienceMetrixLabel: Pharmacology & Pharmacy
journalSubfieldScienceMetrixID: 122
journalSubfieldDepartment: Pharmacology
journalSubfieldScore: 2.19 /* value comes from ScienceMetrixDeptCategory */

If there are multiple matches because a user has multiple departments (infrequent but may occur), choose the highest scoring one.

[jl987-Jie]

Please reference this commit ea3f258 for information on how to query by either eissn or issn.