Whatsout

Project for the Text Mining and Sentiment Analysis exam.

Reported from the description:

Supported by the EPFL Innovation Park, we are working on a new web application helping users to find their favorite activities and events by reducing search effort and improving search relevance. The basic idea to exploit text classification algorithms in order to label events with ad hoc categories. This would allow us to convert unstructured into structured data, thus improving the search relevance and being able to offer a better user experience. Each event has one title, description and author but can be associated to multiple dates, places and categories. The number of categories is not negligible compared to the number of events available in the training set and the dataset is highly imbalanced.

Usage

You can predict the category of an event with one of the trained model.

1. Be sure that the conda/pip environment with the required packages is active.
2. cd into the project folder.
3. python main.py to run it.
4. Choose the model
5. Enter the text of the event.