Logistic Regression in Rental Price and Room Type Prediction Based on Airbnb Open Dataset

Abstract

Based on Aribnb open dataset, this paper is using Logistic Regression—a machine learning method, to analyse how attributes like location and neighbourhood influence the rental price; and, based on the given attributes associate with the estate, predict both rental price and room type. This work is beneficial to the travelers who have the demand in finding an appropriate estate; it can be also instructive in building the recommendation system which can help travelers to find the best estate they want. Apart from the ordinary method in constructing Logistic Regression model which is binary classification, this paper is using softmax function to implement multi-classification which is room type prediction in this work. Through price prediction did not reach the desirable outcome, the room type prediction, however, reached the accuracy about 80%.