Untitled

for i in range(len(data) - window_size - forecast_steps + 1):
    # Select the training data for this fold
    train_data = data.iloc[i:i + window_size]

    # Fit the exponential smoothing model
    model = ExponentialSmoothing(train_data['Value'], trend='add', seasonal='add', seasonal_periods=12)
    fitted_model = model.fit()

    # Forecast for the specified number of steps
    forecast = fitted_model.forecast(steps=forecast_steps)

    # Store the forecasted values
    forecasted_values.append(forecast)

# Convert the list of arrays into a DataFrame
forecast_df = pd.DataFrame(forecasted_values, index=data.index[window_size:window_size + len(forecasted_values)])

# Calculate mean squared error for each forecasted step
mse_per_step = []
actual_values = data['Value'][window_size + forecast_steps - 1:]
for step in range(forecast_steps):
    forecast_values = forecast_df.iloc[:, step]
    mse = mean_squared_error(actual_values, forecast_values)
    mse_per_step.append(mse)

# Print the mean squared error for each forecasted step
for step, mse in enumerate(mse_per_step):
    print(f"Step {step + 1}: Mean Squared Error = {mse:.2f}")