Shows the value which splits the relative error distribution in half. This means that half of the relative errors fall below this value. This value is always positive, since it takes into account the absolute distances between the estimated and the real values. A lower value means better model performance. The optimal value for this measure would be close to 0.