Reza Taleblou
Abstract
In this paper, we evaluate the performance of two machine learning architectures— Recurrent Neural Networks (RNN) and Transformer-based models—on four commodity-based company indices from the Tehran Stock Exchange. The Transformer-based models used in this study include AutoFormer, FEDformer, ...
Read More
In this paper, we evaluate the performance of two machine learning architectures— Recurrent Neural Networks (RNN) and Transformer-based models—on four commodity-based company indices from the Tehran Stock Exchange. The Transformer-based models used in this study include AutoFormer, FEDformer, Informer, and PatchTST, while the RNN-based models consist of GRU and LSTM. The dataset comprises daily observations collected from April 20, 2020, to November 20, 2024. To enhance the generalization power of the models and prevent overfitting, we employ two techniques: splitting the training and test samples, and applying regularization methods such as dropout. Hyperparameters for all models were selected using a visual method. Our results indicate that the PatchTST model outperforms other methods in terms of Root Mean Squared Error (RMSE) for both 1-day and 5-day (1-week) forecasting horizons. The FEDformer model also demonstrates promising performance, particularly for forecasting the MetalOre time series. In contrast, the AutoFormer model performs relatively poorly for longer forecasting horizons, while the GRU and LSTM models yield mixed results. These findings underscore the significant impact of model selection and forecasting horizon on the accuracy of time series forecasts, emphasizing the importance of careful model choice and hyperparameter tuning for achieving optimal performance.
emad koosha; Mohsen Seighaly; Ebrahim Abbasi
Abstract
The purpose of the present research is to use machine learning models to predict the price of Bitcoin, representing the cryptocurrency market. The price prediction model can be considered as the most important component in algorithmic trading. The performance of machine learning and its models, due to ...
Read More
The purpose of the present research is to use machine learning models to predict the price of Bitcoin, representing the cryptocurrency market. The price prediction model can be considered as the most important component in algorithmic trading. The performance of machine learning and its models, due to the nature of price behavior in financial markets, have been reported to be well in studies. In this respect, measuring and comparing the accuracy and precision of random forest (RF), long-short-term memory (LSTM), and recurrent neural network (RNN) models in predicting the top and bottom of Bitcoin prices are the main objectives of the present study. The approach to predicting top and bottom prices using machine learning models can be considered as the innovative aspect of this research, while many studies seek to predict prices as time series, simple, or logarithmic price returns. Pricing top and bottom data as target variables and technical analysis indicators as feature variables in the 1-hour time frame from 1/1/2018 to 6/31/2022 served as input to the mentioned models for learning. Validation and testing are presented and used. 70% of the data are considered learning data, 20% as validation data, and the remaining 10% as test data. The result of this research shows over 80% accuracy in predicting the top and bottom Bitcoin price, and the random forest model’s prediction is more accurate than the LSTM and RNN models.
