-
Twitter has become an important communication channel in times of emergency. The ubiquitousness of smartphones enables people to announce an emergency they’re observing in real-time. Because of this, more agencies are interested in programatically monitoring Twitter (i.e. disaster relief organizations and news agencies).
-
In this competition, you’re challenged to build a machine learning model that predicts which Tweets are about real disasters and which one’s aren’t. You’ll have access to a dataset of 10,000 tweets that were hand classified.
-
I used the BERT model in order to create my own NLP algorithm.
-
I used Optuna to optimize the model.
-
The metric used is "F1-score".
-
The algortihm developed in this Notebook achieved a F1-score of about 82% on the test set (calculated by the Kaggle platform).