Predicting Log Error for House Price Using Machine Learning
Author | : Swapnilkumar Kanani |
Publisher | : |
Total Pages | : 46 |
Release | : 2019 |
ISBN-10 | : OCLC:1236033798 |
ISBN-13 | : |
Rating | : 4/5 (98 Downloads) |
Download or read book Predicting Log Error for House Price Using Machine Learning written by Swapnilkumar Kanani and published by . This book was released on 2019 with total page 46 pages. Available in PDF, EPUB and Kindle. Book excerpt: Buying a house is the major investment decision for an individual. Identifying the property in a desirable location with the right size and build quality at an affordable price is a daunting task. In recent years, many online platforms such as Zillow, Trulia, Redfin, Homesnap have developed algorithms to estimate housing prices across the US. These platforms use statistical model and machine learning algorithm to provide their users the price estimation services at no cost. For example, Zillow has data for 110 million homes across the US. It estimates the price based on the publicly available data including square footage, number of bedrooms, bathrooms, quality of the neighborhood and many other features. The goal of the project is to understand the data analysis techniques and use the understanding to predict the error in the house price prediction. By trying different machine learning algorithms including XGBoost, LightGBM and linear regression, different accuracies can be achieved. The thesis explores four key areas of the machine learning process: Data acquisition, exploratory data analysis, feature engineering, prediction, and validation. Housing data from Orange, Ventura, and Los Angeles counties are used to estimate the log errors. Following formula is used to calculate log error. logerror=log (Prediction)−log (SalePrice)