twosigma_ml

Predicting 'interest level' in NY listings using machine learning

This is a dataset from a competition on Kaggle. The aim is to predict 'interest level' in rental listings. It is a classification problem.

Features explored -

1. Number of bedrooms
2. Number of bathrooms
3. Description
4. Number of photos
5. Location of listing
6. Price point of listing

Based on these features, I trained a LR, Random forest and QDA model. The Random forest model has displayed the best performance.