AI for NLP

• 

  Labels Count

• 

  WordCloud

Twitter has been used as an effective tool to

express their feelings about the disaster event, which can be very helpful for the governing bodies to understand the reactions of the masses and take appropriate actions.

It classifies the sentiments and fine-grained emotions about disaster management from the tweet.

We have built it using the python jupyter notebook and used various python libraries and architectures to get the best accuracy.

As the dataset was not well organized, so we combined all the .csv files and then applied under-sampling techniques to balance the labels.

we have received 91% of accuracy on the test dataset from two architectures. which gives prominent results.

We have learned various machine learning techniques and textual data processing and analysis.

In NLP we can build more complex architecture for Image caption generation, Text to image generation, and Sentiment analysis.

Built With

• google-colab
• python
• tensorflow