Algorithmic trading and price forecasting: machine learning and deep learning strategies for financial markets

This study presents a comparative analysis of various machine learning models for predicting house prices. The models evaluated include linear regression, gradient boosting machine (GBM), k-nearest neighbors (KNN), decision tree, support vector machine (SVM), long short-term memory (LSTM), autoencoder, recurrent neural network (RNN), Bayesian ridge regression, and generative adversarial network (GAN). Evaluation metrics such as root mean squared error (RMSE), mean absolute error (MAE), R-squared, median absolute error (MedAE), and standard deviation (Std) are employed to assess the performance of each model. The results indicate that the autoencoder model achieves the best performance based on the provided metrics, exhibiting exceptionally low RMSE and MAE values along with a high R-squared value. The numerical findings from the comparative analysis of various machine learning models for house price prediction reveal notable differences in performance across different evaluation metrics. The autoencoder model stands out as the top performer, demonstrating remarkably low RMSE and MAE values, both at 0.001 and 0.005, respectively. Additionally, the autoencoder achieves an impressive R-squared value of 0.99, indicating a high level of variance explained by the model. Moreover, the MedAE and Std metrics for the autoencoder are substantially lower compared to other models, further emphasizing its superior predictive accuracy. Conversely, models such as SVM, LSTM, and GAN exhibit comparatively poorer performance, as indicated by higher RMSE, MAE, and lower R-squared values. Overall, these findings underscore the effectiveness of the autoencoder model for house price prediction tasks.