GitHub - Doffery/BERT-Sentiment-Analysis-Amazon-Review

#Bert Amazon Review Sentimental Analysis

A project leveraging BERT and Amazon Review Data to do sentimental analysis.

We tried fine-tuned bert (run_classifier_n.py), ngram as baseline (ngram_baseline.py) and combining both feature set (hybrid_ar_sa.py).

Name		Name	Last commit message	Last commit date
Latest commit History 6 Commits
BertVectors		BertVectors
.gitignore		.gitignore
CONTRIBUTING.md		CONTRIBUTING.md
LICENSE		LICENSE
README.md		README.md
__init__.py		__init__.py
arrun.sh		arrun.sh
computeBERTSentenceEmbedding.py		computeBERTSentenceEmbedding.py
conda-bert-env-win.yaml		conda-bert-env-win.yaml
conda-bert-env.yaml		conda-bert-env.yaml
create_pretraining_data.py		create_pretraining_data.py
data_formating.py		data_formating.py
extract.sh		extract.sh
extract_features.py		extract_features.py
extract_features_m.py		extract_features_m.py
extract_features_n.py		extract_features_n.py
fcnn_baseline.py		fcnn_baseline.py
hybrid_ar_sa.py		hybrid_ar_sa.py
input.txt		input.txt
local_bert.py		local_bert.py
modeling.py		modeling.py
modeling_test.py		modeling_test.py
multilingual.md		multilingual.md
ngram_baseline.py		ngram_baseline.py
optimization.py		optimization.py
optimization_test.py		optimization_test.py
requirements.txt		requirements.txt
run_classifier.py		run_classifier.py
run_classifier_n.py		run_classifier_n.py
run_pretraining.py		run_pretraining.py
run_squad.py		run_squad.py
sample_text.txt		sample_text.txt
tokenization.py		tokenization.py
tokenization_test.py		tokenization_test.py
utils.py		utils.py

Provide feedback