Modern measurement techniques like scanning technology or sonar measurements, provide large datasets, which are a reliable source of information about measured object, however such datasets are sometimes difficult to develop. Therefore, the algorithms for reducing the number of such sets are incorporated into their processing. In the reduction algorithms based on the cartographic generalization method, it is required to input some parameters (e.g. tolerance), which are determined by the user. The choice of the values of parameters, and in results the number of points in the reduced set, is one of the key step in the algorithm's efficiency. Thus, it requires from the user to have the knowledge on how the reduction algorithm works, and what is the relationship between the values of these parameters and the final number of points in reduced set. In this article authors used the regression analysis to explore this aspect of processing the large datasets.