Comparative Analysis of Decision Tree, Neural Network, K-NN, GBT, SVR and Random Forest Algorithms for House Price Prediction

Sahal Abdillah (Universitas Nahdlatul Ulama Al-Ghazali Cilacap)
Mochamad Taufiqurrohman Abdul Aziz Zein (Universitas Nahdlatul Ulama Al-Ghazali Cilacap)
Edy Sulistiyanto (Universitas Nahdlatul Ulama Al-Ghazali Cilacap)

COREAI: Jurnal Kecerdasan Buatan, Komputasi dan Teknologi Informasi
Vol 7, No 1 (2026): Sustainable Information Technology Innovation Supports a Digital-Based Smart Eco

This study aims to analyze and compare the performance of several machine learning algorithms for predicting house prices. The dataset was obtained from Kaggle with variables of house price, building area, land area, number of bedrooms, number of bathrooms, and garage. The algorithms used include Decision Tree, Neural Network, K-Nearest Neighbor, Gradient Boosted Trees, Support Vector Regression, and Random Forest. The study was conducted using Altair Studio through pre-processing, training, testing, and evaluation stages based on performance, computation time, and model complexity. The results showed that K-NN obtained the best RMSE value of 0.608, followed by Neural Network 0.630, Decision Tree 0.649, Random Forest 0.656, SVR 0.662, and GBT 0.747. Based on MAE, Neural Network obtained the best value of 0.464, followed by Random Forest 0.474, Decision Tree 0.478, K-NN 0.485, SVR 0.501, and GBT 0.589. Overall, K-NN was the best model because it had the lowest RMSE, the fastest training time of 0.004 seconds, and a relatively simple model complexity.