The behavior of residuals of linear on parameters regression MLS-model with the increase of the number of parameters Pt.2 the interval for the residuals series. the problem of crude errors

摘要

The distribution of residuals in one series of measurements is obtained. It's shown, that the t-interval for residuals, which is usually used, is always wider than the correct one. The question of gross errors detection on residuals' plots is discussed. The reliable crude errors detection is possible only when the ratio of the number of parameters of the model to the number of points is very small (≤0,1).