The modeling objective was to predict the number of inquiries a new listing receives based on the listing’s creation date and other features such as text description, photos, number of bedrooms, price, etc.

 

• Constructed the Xgboost model and used cross-validation and used Bayesian Optimization to tune parameters

 

• Formulated powerful features using techniques such as TFIDF, Countvectorizer, and features crossing

 

Github: https://github.com/StephanieZ001/Kaggle-twosigma/blob/master/TwoSigma2.ipynb