- Here we will predict the emotions from dialogues using Machine Learning and Deep Learning techniques
- We will create a very basic first model first and then improve it using different other features. We will also see how deep neural networks can be used and end this post with some ideas about ensembling in general.
- TFIDF
- BOW
- logistic regression
- naive bayes
- svm
- xgboost
- grid search
- word vectors
- LSTM
- Customised Ensembling
Research PAPER: https://arxiv.org/pdf/1810.02508.pdf
DATASET : MELD : https://affective-meld.github.io/
- Firstly, load all the dependencies.
- I have kept things simple and consolidated into single .ipynb file.
- Just go to the end of 'EmotionfromConv.ipynb' and run the Inference model and get the desired output.
- To load Glove file : https://drive.google.com/file/d/1zA-Tc6pnFvXxnGK54-aUxBqmlV2y50mH/view?usp=sharing