ELJST-Paper

Instructions to run a new dataset.

Make a copy of ELJST_Onetime and import the dataset and make the necessary preprocessings.
Use the same ipynb to generate the edges.
further, Test_ELJST_Onetime_only_edges.ipynb and Test_ELJST_Onetime_only_edges_attention_all.ipynb can be used to generate edges.
attention_all can be generated only in attention_all.ipynb.
all the edges will be in the resources folder.
baselines can be run using the nohookup (nosampler_uni, nosampler_etm (mrf-lda), for etm there is a different repo).
for lda, Baseline_LDA ipynb. Evaluations for LDA will be in the same file
once the models are dumped, evaluation scipts should be run seperately.
For MRF-LDA: Evaluation_ETM
For ELJST: ELJST-Evaluation
For ETM: Seperate Repo.

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
.ipynb_checkpoints		.ipynb_checkpoints
datasets		datasets
wordvecs		wordvecs
.gitignore		.gitignore
Baseline_ETM.ipynb		Baseline_ETM.ipynb
Baseline_ETM_Evaluations.ipynb		Baseline_ETM_Evaluations.ipynb
ELJST-Evaluation.ipynb		ELJST-Evaluation.ipynb
ELJST_script_unigram.py		ELJST_script_unigram.py
JST.py		JST.py
LDA_ETM.py		LDA_ETM.py
NMF_Topic_Sentiment_Modeling.pdf		NMF_Topic_Sentiment_Modeling.pdf
README.md		README.md
RJST.py		RJST.py
TSM.py		TSM.py
TSWE.py		TSWE.py
Test_ELJST_Onetime_only_edges.ipynb		Test_ELJST_Onetime_only_edges.ipynb
Test_ELJST_Onetime_only_edges_attention_all.ipynb		Test_ELJST_Onetime_only_edges_attention_all.ipynb
Train_Embedding_with_sentiment.ipynb		Train_Embedding_with_sentiment.ipynb
Train_Embeddings_Fasttext.ipynb		Train_Embeddings_Fasttext.ipynb
__init__.py		__init__.py
base.py		base.py
noedges.py		noedges.py
requirements.txt		requirements.txt
sLDA.py		sLDA.py
utils.py		utils.py
utils_tswe.py		utils_tswe.py