To choice the best regression tree method, goodness of fit criteria such as coefficient of determination (R2%), adjusted coefficient of determination (Adj-R2%), coefficient of variation (%), SD ratio, relative

approximation error (RAE), Root Mean Square Error (RMSE), Pearson correlation between actual and predicted weaning weights were estimated for each combination.

An estimation of the

approximation error, as well as the testing of the accuracy of the method by considering distribution of the

approximation error for experimentally obtained SHLC, is presented in Section III.

We quantify the algorithm performance by measuring the

approximation error, compression ratio, and computation complexity.

i] is neural network

approximation error that satisfies [absolute value of ([[epsilon].

2 (Table 2) shows that the basic ensemble approach allows us to obtain an

approximation error of magnitude close to the magnitude of uncertainty in the data.

On the other hand, in order to get the same

approximation error, the proposed method will require significantly fewer samples than uniform sampling.

i]} - set of adjusting parameters of MCF (the left and right borders of the cut-out interval and

approximation error setting e.

Thus, Table 2 shows that a slightly smaller

approximation error for this software can be achieved using activation function Inverse Multiquadric and configuration "5-10" or "15-10" (2,4% and 3,0% accordingly), but the training duration is 1618 and 1536 periods against 324 periods in the case of configuration "10-30".

This is also the module responsible for most of the final

approximation error as it aggregates data from different processing paths.

Based on the weighted-residual error estimator from [7], we introduced an overall error estimator which controls both, the discretization error as well as the data

approximation error (Theorem 3.

It is shown that the

approximation error increases as the frequency decreases.

In such a case, the Taylor's theorem ensures that the remainder term e, that is the

approximation error given by the difference between the real value attained by the function and its Taylor polynomial, is negligible if compared to the size of [(x - a).