Skip to content

Academic Paper/Document classifier based on SCOPUS categories

Notifications You must be signed in to change notification settings

sammous/allenNLP_tagger

Repository files navigation

Academic Paper/Document classifier based on SCOPUS categories

Usage

Paper classifier taking as input a JSON with follow fields :

text: string

And returning probabilities for over 237 categories extracted from SCOPUS.

## Training

To train a model :

python -m allennlp.run train allen_trainer/scopus_classifier.json \
    -s model \
    --include-package papers

Making Predictions

Making predictions on a JSON file :

allennlp predict model/model.tar.gz \
    scopus_test.json \
    --include-package papers \
    --predictor paper-classifier \
    --output-file scopus_test_pred.json

## Demo

Visualisation demo for a text classifier.

The folder demo contains the front app to demo a model trained via allenNLP with some visualisations. Examples can be previewed on their demo website.

Requirements

allenNLP server must be running with a classification model (allenNLP simple server) on port 8000.

Example command :

python -m allennlp.service.server_simple \
    --archive-path model/model.tar.gz \
    --predictor paper-classifier \
    --include-package tagger 

Running demo

To run the demo :

cd demo
sudo npm install
sudo npm start

Credits

About

Academic Paper/Document classifier based on SCOPUS categories

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published