HOUSE PRICE PREDICTION USING LINEAR REGRESSION

Abstract

Prediction of prices of houses is very important for buyers, sellers, and planners in the real estate market. In this paper, it is explained how Linear Regression can be used for estimating and predicting the prices of houses using the Ames Housing Dataset. Ames Housing Dataset includes data of 1460 houses and their 81 different features or attributes about the houses. We first cleaned missing data of the dataset using simple methods like replacing them with median or most common values. Then we trained the Linear Regression Model and tested it. Our model resulted in 0.82 score for R², 22417 for MAE, and an RMSE of about 36,879. From this, we understand that even a simple model like Linear Regression can work well in predicting prices of houses. In the end, we also discussed various studies related to our work and also included how we can improve this system in the future.