Web-deployment : https://nyc-rental-price.herokuapp.com/
Stayze is an online market for providing lodging or primary homestays. The company does not own any real estate or properties, it acts as a broker receiving commission from each booking. The hosts rent out their property, its availability, area, type of room, price etc. and the travellers can book accordingly. The travellers put in their reviews, which is visible to others. People have used this service extensively & the company is recognized throughout the globe. All the online activities of the hosts as well as the travellers are being captured and have resulted in a rich database. This data can be used to gain business insights, make decisions, improve security, understand the customers' and providers' (hosts) behaviour and performance on the platform, guiding marketing initiatives, implementation of innovative additional services and much more. The stakeholders with the help of the available data want to know the ideal prices at which the properties can be rented, as it will help them decide upon the ideal investment to be done. Build a machine learning model to predict the ideal price of the rental.
Train.csv - It is the training data containing the features, along with the price of the rentals. Data_Dictionary.xlsx - It contains a brief description of every variable provided in the training and test set. Test.csv: - It contains details of the customers for which the participants need to predict the price of the rentals.
For evaluation, a solution with low root-mean-squared error (RMSE) was desired.
Data Pre-processing, Data Visualization, Outlier Analysis, Missing value Analysis, Feature Selection, Correlation analysis, Multicollinearity Test, Feature Scaling, Splitting into Train & Test Dataset, Hyper-parameter Tuning.
I. Ridge Regression II. Random Forest III. XGBoost