The aim of this paper is to explain the choices and the strategies we adopted on the project and development of AirBnb Price Estimator, whose aim is to help owners to decide the most correct price for their BnB. In order to accomplish it, we started from web-scraped data, we performed all the preprocessing needed for having a suitable dataset and then we built several classifiers, using different strategies, in order to determine the one that predicts best the class attribute. All these classifiers have been tested using more than one method and the analysis of the results guided us in the choice of the best classifier.
Classification algorithms used
- Linear Regression w/o attribute selection
- Random Forest w/o attribute selection
- k-NN (k=5) w/o attribute selection
- M5Rules w/o attribute selection
Tech Stack Adopted
- JavaFX and Java 8
- Weka
Main Libraries
- Weka API
