Streamlining the peer review process
Expert Finder is a web-accessible tool designed to aid academic professionals in the peer review process by querying scholarly libraries to compile a list of experts for a specific topic.
Topic Search is designed to help the peer review process for a single academic manuscript. Enter a manuscript's abstract and title to recieve a list of academic professionals that have similar fields of research, ranking them based on how similar thier work is to the manuscript.
- Rank potential reviewers based on number of papers written on the searched subject, the number of times their papers have been cited, and the time since the reviewer has published a paper related to the topic
Author search allows the user to find potential reviewers for an already established author. The user queries the authors name, from there the user is allowed to select a subset of the authors papers. After doing so, a list of ranked reviewers is returned to the user. Each ranked based on their ability to review the author's body of work.
Each author is ranked based on the average of a weighted metric calculated for each paper. The document score for each of an authors papers is calculated as a weighted sum of document similarity, number of citations, and recency of the paper. Similarity is weighted highest, followed by citations then recency. An author’s final score is the average document score of their ten highest scoring papers (average of all paper scores for authors with less than ten).
Cosine Similarity is used to measure how similar a given document is to the user’s query. Each document is represented as a vector composed of its TF-IDF (Term Frequency-Inverse Document Frequency) scores with respect to the terms in the query. The TF-IDF score of a word indicates how important that word is to the document that contains it when compared to the entire list of documents. The cosine of the angle between the TF-IDF vectors represents how similar the documents are.
ExpertFinder is powered by the Django Web framework
ExpertFinder uses the academic databases of Microsoft Academic to provide its service. Microsoft's Academic Knowledge API provides access to its academic databases through the REST API.
ExpertFinder takes advantage of memcached's caching service and uses it to cache the results that it retrieves from Microsoft Academic. Doing so allows ExpertFinder to reduce the overall number of calls it must make to the database and allows it to provide a faster service.
- Django==1.10.1
- nltk==3.2.1
- numpy==1.11.2
- python-memcached==1.58
- requests==2.11.1
- six==1.10.0
- textblob==0.11.1
- virtualenv==15.1.0
- Download code
- Install python dependencies
- pip install -r requirements.txt
- Download nltk stopwords
- python
- import nltk
- nltk.download('stopwords')
- Download corpus
- python -m textblob.download_corpora
- Run the development server
- python manage.py runserver
- https://docs.djangoproject.com/en/1.10/intro/tutorial01/
- https://docs.djangoproject.com/en/1.10/
- https://docs.djangoproject.com/en/1.10/intro/overview/
- https://www.microsoft.com/cognitive-services/en-us/Academic-Knowledge-API/documentation/EvaluateMethod
- https://www.microsoft.com/cognitive-services/en-us/Academic-Knowledge-API/documentation/InterpretMethod
- https://www.microsoft.com/cognitive-services/en-us/Academic-Knowledge-API/documentation/QueryExpressionSyntax
- https://dev.projectoxford.ai/docs/services/56332331778daf02acc0a50b/operations/565d753be597ed16ac3ffc03
- https://dev.projectoxford.ai/docs/services/56332331778daf02acc0a50b/operations/56332331778daf06340c9666/console