Predicting Future Price of ۱۸ Carat Gold in Iran: A Long Short-Term Memory-Based Approach

With the escalating value of ۱۸ carat gold in Iran, the imperative need for a price prediction model becomes evident to facilitate informed investment decisions. This article introduces a novel approach utilizing Long Short-Term Memory (LSTM) for forecasting future prices of ۱۸ carat gold in Iran. The article begins by examining various types of kernels for support vector regression (SVR) to determine the optimal kernel for predicting the price of ۱۸ carat gold in Iran. Subsequently, it compares the predictive performance of the LSTM model with SVR. Experimental results reveal the LSTM neural network's superior predictive accuracy compared to SVR. The study proposes a one-day forecast range for ۱۸ carat gold prices in Iran, emphasizing the LSTM-based model's capacity to provide precise predictions. In conclusion, this paper firmly establishes that LSTM, recognized for its exceptional performance in tackling complex prediction tasks, consistently outperforms traditional regression-based models.

Keywords:

Forecast , ۱۸ Carat Gold , Recurrent Neural Network , Long Short-Term Memory , Support vector regression