Real Estate Price Prediction using Linear Regression

Click to View Rpubs Report
Click to View Github Code and Dataset

OBJECTIVE

This Machine Learing Project involves Building a Predictive model using Linear Regression and predicting the accurate prices of the Real Estate proterties.

Price of a property is one of the most important decision criterion when people buy homes. Real state firms need to be consistent in their pricing in order to attract buyers . Having a predictive model for the same will be great tool to have , which in turn can also be used to tweak development of properties , putting more emphasis on qualities which increase the value of the property.

TAGS

Linear Regression, RMSE.

PROJECT METHODOLOGY

The project involves the following steps:

• 1. Imputing NA values in the datasets.
• 2. Data Preparation:Grouping similar category variables and making dummies.
• 3. Model Building. ( LINEAR REGRESSION )
• 4 .Perfomance measurement of the model. (Calculating RMSE)
• 5. Predicting Real Estate Prices for the final Test Dataset.

DATA DICTIONARY

• There exist two datasets, housing_train.csv and housing_test.csv .
• We will use data housing_train to build predictive model for response variable “Price”.
• Housing_test data contains all other factors except “Price” which we can use for testing purpose.
• The evalution metric will be RMSE.

CONCLUSION

Real estate price prediction was done successfully using linear regression model having Adjusted R-square: 0.6764.

Real (test_25$Price) vs Predicted (PP_test_25) Plot

Residual Vs Fitted Plot (A test for Linearity)

Q-Q Plot (for Normality)

Cooks Distance Plot( for knowing Outliers)