House Price Prediction

Abstract

Property is one of necessities of human life. The demand for homes increased quickly over time as people’s living standards rose. One of the important tasks in real estate is to predict the price of property. The customers don’t have a full background history of the property which they intend to buy or sell so most of them end up getting scammed. And the customers don’t know whether the property will gain or loss in future. So, we build up a predictive model for these problems using machine learning technique. The technique which gave the better accuracy was Random Forest regressor. So, data from Zameen.com was trained on this technique to predict the price. For the front-end Django web-application is made. This model is for anyone who is buying or selling a property especially for the user who wants to predict the price for particular city. At the end very user-friendly application is made for the users. In Pakistan Zameen.com and Graana.com are very popular platforms but neither of them has a predicted model. They show trends but cannot really provide efficient housing price model. ML played a crucial role in our project. We first used various regression algorithms, including linear regression and decision tree regression but the model which we used in framework is Random Forest regression because it provides better accuracy. Through iterative testing and evaluation, we discovered that random forest regression model yielded the best predictive performance for house prices. This experience taught us that the choice of right algorithm can significantly influence the success of a predictive model.