Differentiate between mis/dis/mal-information in news using NLP to track actors and their interconnectivity with each other and world evens in a holistic fashion.
Love ahead by a nose.
This is using the same library as flair-news in my shed.
I've uploaded sample.cypher.gz so you can play along at home.
This took an entire day to load. You might want to try head sample.cypher to make sure it works.
1505 news articles
ready to start consuming query after 18691584 ms
Added 28877 nodes, Created 144549 relationships, Set 320810 properties, Added 28768 labels
- Gunzip
gunzip sample.cypher.gz- The user is set to neo4j, set the password
export NEO4J_PASS='state your password'- Run the docker container.
make neo4j- Load the sample. This might literally take all day.
make sampleloadThis query will return the top 10 nodes with the highest bias, along with information about their LOVES and HATES relationships.
MATCH (n)-[r:BIAS]->(m)
OPTIONAL MATCH (n)-[loves:LOVES]->(lovedNode)
OPTIONAL MATCH (n)-[hates:HATES]->(hatedNode)
RETURN n, r.bias AS biasScore,
collect(DISTINCT {type: 'LOVES', target: lovedNode}) AS lovesRelationships,
collect(DISTINCT {type: 'HATES', target: hatedNode}) AS hatesRelationships
ORDER BY r.bias DESC
LIMIT 10MATCH (n)
WITH n, COUNT{(n)--() } AS connectionCount
RETURN n, connectionCount
ORDER BY connectionCount DESC
LIMIT 10MATCH (n:PERSON)
WITH n, COUNT{ (n)--() } AS connectionCount
RETURN n, connectionCount
ORDER BY connectionCount DESC
LIMIT 10MATCH (n:WORK_OF_ART)
WHERE n.val IN ["Beetlejuice Beetlejuice", "Beetlejuice"]
MATCH (n)-[r]-(connected)
RETURN n, r, connectedMATCH (n:PERSON)
WITH n, COUNT{ (n)--() } AS connectionCount
RETURN n, connectionCount
ORDER BY connectionCount DESC
LIMIT 20MATCH (n)-[r:LOVES|HATES]->(m) RETURN n, type(r), mThis will return the source node, the relationship type (LOVES or HATES), and the target node.
create user kyle if not exists set plaintext password "stupidpassword" change not requiredMATCH (n) DETACH DELETE nbin/cypher-shell -u neo4j -p "password" -f scripts/movies.cypher [-d "database"]MATCH (n) MATCH ()-[r]->() RETURN n, rMATCH (n) RETURN (n)MATCH (n) RETURN n UNION START n = rel(*) return n;flair framework:
@inproceedings{akbik2019flair,
title={{FLAIR}: An easy-to-use framework for state-of-the-art {NLP}},
author={Akbik, Alan and Bergmann, Tanja and Blythe, Duncan and Rasul, Kashif and Schweter, Stefan and Vollgraf, Roland},
booktitle={{NAACL} 2019, 2019 Annual Conference of the North American Chapter of the Association for Computational Linguistics (Demonstrations)},
pages={54--59},
year={2019}
}
@InProceedings{Hamborg2021b,
author = {Hamborg, Felix and Donnay, Karsten},
title = {NewsMTSC: (Multi-)Target-dependent Sentiment Classification in News Articles},
booktitle = {Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2021)},
year = {2021},
month = {Apr.},
location = {Virtual Event},
}