cARM

create Astro Ratings Matrix

cARM is a software tool that creates a dataset suitable for evaluating recommender systems for Open Cluster of Stars using Scientific Literature.

Dependencies:

• python 3.5+
• sqlite
• numpy
• pandas
• configargparse

Configuration steps

• The first step is to create an account at ADS - follow the instructions in https://github.com/adsabs/adsabs-dev-api to create your ADS API token

• Second, you need to create the sqlite database - the commands are in createDB.sql

• Third, make sure to complete the info at config.ini file, including the correct paths to the database and clusters file

RUN

• Now you need to fill the database with the information about the clusters:

python fillClustersTable.py

• Next step is performed to find all the articles for each cluster in our database

python fillTables.py 

Create <user,item,rating> dataset

cd authorClusterMatrix/
python createAuthorClusterMatrix.py

Create sequential dataset

cd authorClusterMatrix/
python createAuthorClusterMatrix_seq.py

Note:

_{^{The script /clustersNameAnalysis/clustersNameAnalysis.py returns groups of names of clusters whose simbadID was not found}}

Dataset

For download of an example of dataset created with this software: https://drive.google.com/drive/folders/17a5rSq1iIy4vsgAuj0TGHTY3UbaVpvqq?usp=sharing

Here is a simple script to create the subsets of the full dataset, by removing users with a minimal number of rated items: https://bitbucket.org/snippets/mcabarros/aenGGo/remove_users_by_min_rating