TOP 150 (TOP14%) Kaggle on 18th of february 2021
- Link of the Competition - https://www.kaggle.com/c/nlp-getting-started
Real or Not? NLP with Twitter.
As a Datascience newcomer, I decided to write this notebook to extend my knowledge about natural language processing (NLP), in this notebook, we will be analyzing 10,000 hand classified tweets. these tweets are either labeled as disasters or not disasters. For instance we can consider the 2 following tweets:
- These boxes are ready to explode! Exploding Kittens finally arrived! (Not Disaster)
- Forest fire near La Ronge Sask. Canada (Disaster)
So this kernel aim's at predicting the correct label. We will go through data analysis, cleaning and text processing, feature engineering, word embedding with glove and we will finally make 3 differents DL models:
- Simple Neural Network with text input
- Mixed Data Neural Network with text input + metadata input
- Neural Network with Bert pretrained layer
Please note that this is my first notebook published on Kaggle, any advice would be highly appreciated, do not hesitate if you have any idea to improve this kernel, feel free to ask questions aswell.