House Price Prediction using Random Forest Machine Learning Technique

Predicting a price variance rather than a specific value is more realistic and attractive in many real-world applications. Price prediction can be thought of as a classification issue in this situation. However, the House Price Index (HPI) is a common tool for estimating the inconsistencies of house prices. Since housing prices are closely correlated with other factors such as location, city, and population, predicting individual housing prices needs information other than HPI. The HPI is a repeat-sale index that tracks average price shifts in repeat transactions or refinancing of the same assets. Therefore, HPI is ineffective at predicting the price of a single house because it is a rough predictor based on all transactions. This study explores the use of Random Forest machine learning technique for house price prediction. UCI Machine learning repository Boston housing dataset with 506 entries and 14 features were used to evaluate the performance of the proposed prediction model. A comparison of the predicted and actual prices predicted revealed that the model had an acceptable predicted value when compared to the actual values with an error margin of ±5.