README

This repository is a university project showcasing my NLP (Natural Language Processing) skills and demonstrating the use of BERT (Bidirectional Encoder Representations from Transformers).

Airbnb Property Unlisting Prediction

The goal of this project is to utilize Natural Language Processing (NLP) models to predict whether a property listed on Airbnb will be unlisted in the next quarter. To accomplish this, we will leverage real Airbnb property descriptions, Airbnb host descriptions, and comments from previous guests.

Using the NLP techniques, we will implement an NLP classification model that can accurately predict, for each property, whether it will be unlisted (1) or remain listed (0).

Notebooks

There are two different notebooks available in this repository:

1. notebook_short.ipynb: This notebook is visible directly on GitHub due to its smaller file size. It contains the essential code and explanations.

2. notebook_long.ipynb: This notebook is not visible on GitHub due to its larger file size. However, you can download it and examine it locally to explore the complete project in detail.

Files

The repository includes the following important files:

• .gitattributes: Configuration file for Git attributes.
• NLP_16.csv: Dataset file containing the predictions on the test data.
• test.xlsx: Additional test data file.
• test_reviews.xlsx: Test data file for reviews.
• train.xlsx: Additional training data file.
• train_reviews.xlsx: Training data file for reviews.

Please note that only the essential files are included in the repository due to their large size. However, you can find everything you need within the notebooks themselves.

Feel free to explore the notebooks and the provided data to gain insights into my NLP skills and the implementation of BERT.