Emotion Detection using Trained LSTM

Implementation with sad emotion
Implementation with happy emotion

Inspiration

We have always been thinking about whether an AI can detect emotions in some type of writing or typing and the application of it.

What it does

The trained LSTM detects emotions from the text provided to it.

How we built it

We used the Jupyter Notebooks with Tensorflow, and scikit learn libraries.

Challenges we ran into

The model misclassifies the emotions in the text.

Accomplishments that we're proud of

Instead of using a large language model that requires extensive training, our model performs pretty well at a 93 percent accuracy

What we learned

We learned how Neural Nets represent and process words and LSTMs can only detect emotions well if the emotions words are used in the sentence. We also learned how to implement a virtual cloud platform via ssh connection. This hackathon has also taught us teamwork and the mindset to overcome adversities.