Performed and compared various linear regressions techniques (penalize, LASSO and Ridge), Generalized Additive Models(GAMs) and Regression trees on the Boston Airbnb dataset to investigate the following questions:
-
To what degree can supervised machine learning techniques be used to assist an Airbnb host in determining an appropriate listing price for their property?
-
For Airbnb data, can the categorical feature of “neighborhood” be replaced with a con- tinuous feature of driving distance to a geographic point of interest (e.g., an airport) and have comparable results?
The complete investigation and report is in documents/final_report.pdf
The R source code is in src/boston_airbnb.Rmd
- Nakul Camasamudram
- Philip Parker
- Abhay Kasturia