COMPARISON BETWEEN CLSSICAL MODELS (VAR& VECM) & A.I(ANN & RNN INCLUDING LTSM & GRU) MODELS.

This Python script performs a comprehensive comparison of classical time series models and AI-based models on synthetic financial datasets. It generates synthetic gold price data (Open, Close, High, Low) and banking data (Interest Rate, Deposits, Loans, Exchange Rate) for 200 daily observations. The script then:

Applies classical models – VAR (Vector Autoregression) and VECM (Vector Error Correction Model) – to analyze relationships between gold prices and banking indicators and to forecast future prices.

Implements AI models – ANN (Artificial Neural Network), Simple RNN, LSTM, and GRU – to predict gold prices based on bank data sequences.

Evaluates model performance using RMSE and R² metrics.

Plots the actual gold prices, AI predictions, and VAR/VECM forecasts in a single figure for direct visual comparison.

The script is written in a single-page format, making it easy to run and modify. It demonstrates the effectiveness of classical econometric models versus AI sequence models in financial forecasting.

Author: Prof. Eldirdiri Fadol Ibrahim