Study Regarding the Spline Interpolation Accuracy of the Experimentally Acquired Data

Experimental data processing is an issue that must be solved in almost all the domains of science. In engineering we usually have a large amount of data and we try to extract the useful signal which is relevant for the phenomenon under investigation. The criteria used to consider some points more relevant then some others may take into consideration various conditions which may be either phenomenon dependent, or general. The paper presents some of the ideas and tests regarding the identification of the best set of criteria used to filter the initial set of points in order to extract a subset which best fits the approximated function. If the function has regions where it is either constant, or it has a slow variation, fewer discretization points may be used. This means to create a simpler solution to process the experimental data, keeping the accuracy in some fair good limits.