Root Mean Square Error Calculation in Matlab

What is Root Mean Square Error (RMSE)?

Root Mean Square Error (RMSE) is a measure of the differences between predicted values and observed values. It provides a single number that represents the average magnitude of the errors between predicted and actual values. RMSE is widely used in regression analysis, forecasting, and machine learning to evaluate model performance.

RMSE is particularly useful because it penalizes larger errors more heavily than smaller errors, making it sensitive to outliers. This makes it a good choice when you want to ensure your model doesn't have large prediction errors.

RMSE Formula

The formula for RMSE is derived from the standard deviation of the residuals (the differences between predicted and actual values). Here's the mathematical representation:

This formula calculates the square root of the average of the squared differences between predicted and actual values. The square root ensures that the units of RMSE match the units of the original data.

MATLAB Implementation

MATLAB provides several ways to calculate RMSE. You can use built-in functions or write custom code. Here are two common approaches:

Using the rmse Function

MATLAB's Statistics and Machine Learning Toolbox includes a built-in rmse function that simplifies the calculation:

% Example data
actual = [3, 5, 2, 7, 9];
predicted = [2.5, 5.1, 1.8, 6.9, 8.8];

% Calculate RMSE
error = rmse(actual, predicted);

% Display result
fprintf('RMSE: %.4f\n', error);

Custom RMSE Calculation

If you don't have the Statistics and Machine Learning Toolbox, you can calculate RMSE manually using basic MATLAB functions:

% Example data
actual = [3, 5, 2, 7, 9];
predicted = [2.5, 5.1, 1.8, 6.9, 8.8];

% Calculate squared differences
squared_errors = (actual - predicted).^2;

% Calculate mean of squared errors
mean_squared_error = mean(squared_errors);

% Calculate RMSE
rmse_value = sqrt(mean_squared_error);

% Display result
fprintf('RMSE: %.4f\n', rmse_value);

Visualizing RMSE with MATLAB

You can visualize the RMSE calculation by plotting the actual vs. predicted values and the error terms:

% Example data
actual = [3, 5, 2, 7, 9];
predicted = [2.5, 5.1, 1.8, 6.9, 8.8];

% Calculate RMSE
rmse_value = sqrt(mean((actual - predicted).^2));

% Create figure
figure;
subplot(2,1,1);
plot(actual, 'b-o', 'DisplayName', 'Actual');
hold on;
plot(predicted, 'r--s', 'DisplayName', 'Predicted');
legend('Location', 'best');
title('Actual vs Predicted Values');
xlabel('Observation');
ylabel('Value');

subplot(2,1,2);
stem(actual - predicted, 'k-o');
title('Error Terms');
xlabel('Observation');
ylabel('Error');

% Display RMSE
annotation('textbox', [0.2, 0.1, 0.6, 0.1], 'String', ...
    sprintf('RMSE: %.4f', rmse_value), 'FitBoxToText', 'on', ...
    'BackgroundColor', 'white', 'EdgeColor', 'none');

Example Calculation

Let's walk through a practical example to calculate RMSE in MATLAB. Suppose you have the following actual and predicted values for a set of observations:

Using the custom RMSE calculation code:

The RMSE for this example is approximately 0.286, indicating that the model's predictions are, on average, about 0.286 units away from the actual values.

Interpreting RMSE Results

Interpreting RMSE requires understanding the context of your data and the scale of your measurements. Here are some guidelines:

Scale of RMSE

RMSE is in the same units as the original data. For example, if you're predicting house prices in dollars, an RMSE of $20,000 means predictions are, on average, $20,000 off from actual prices.

Comparing Models

RMSE is useful for comparing different models. A lower RMSE indicates better model performance. However, always consider other metrics and the context of your problem.

RMSE vs. MAE

RMSE is more sensitive to outliers than Mean Absolute Error (MAE). If your data has outliers, RMSE will be more affected by them. In such cases, MAE might be a better metric.

Perfect vs. Imperfect Models

A perfect model would have an RMSE of 0, meaning all predictions match the actual values exactly. In practice, models will have some error, and RMSE helps quantify how much error exists.