The confidence interval is based on mean square error. The mean squared error (MSE) of an estimator measures the average of the squares of the errors or deviations. MSE is calculated by the formula attached in the picture, where Xbar is a vector of predictions, X is the vector of predicted values.
