Real estate price forecasting utilizing recurrent neural networks incorporating genetic algorithms

Ting Tin Tin, Cheok Jia Wei, Ong Tzi Min, Boo Zheng Feng, Too Chin Xian

Abstract

This study aims to develop and examine the effectiveness of the Recurrent Neural Network (RNN) model incorporating Genetic Algorithm (GA) in forecasting real estate prices. Real estate prices have a significant impact on a country’s financial system. Therefore, the ability to accurately forecast its price is valuable. A set of data containing 5.4 million unique records of real estate with their prices is used in the study. The data set, which spans from 2018 to 2021, contains twelve independent variables and one dependent variable. We preprocessed the data set to reduce noise and outliers that could potentially lead to poor model performance. The RNN model was selected because (GA) optimises the hyperparameters in the Recurrent Neural Network (RNN) hidden layer to optimise the performance of the RNN model. Relative Root Mean Squared Error (RRMSE), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and Mean Absolute Error (MAE) examine the effectiveness of the RNN-GA model.The results show that the RNN-GA outperforms RNN and the traditional statistical models used in the real estate industry. This study provides an understanding of the effectiveness of incorporating GA in optimizing the RNN model for real estate price forecasting, which could benefit stakeholders in the real estate industry and sustain the financial system. This study uses a novel hybrid RNN-GA model for predicting housing prices with a large dataset.

Authors

Ting Tin Tin
tintin.ting@newinti.edu.my (Primary Contact)
Cheok Jia Wei
Ong Tzi Min
Boo Zheng Feng
Too Chin Xian
Tin, T. T. ., Wei, C. J. ., Min, O. T., Feng, B. Z. ., & Xian, T. C. . (2024). Real estate price forecasting utilizing recurrent neural networks incorporating genetic algorithms. International Journal of Innovative Research and Scientific Studies, 7(3), 1216–1226. https://doi.org/10.53894/ijirss.v7i3.3220

Article Details

No Related Submission Found