The paper presents the mini-models' method (MM-method) based on n-dimensional simplex. Its learning algorithm is in some respects similar to the method of k-nearest neighbors. Both methods use samples only from the local neighborhood of the query point. In the minimodel method, group of points which are used in the model-learning process is constrained by a polytope (n-simplex) area. The MM-method can on a defined local area use any approximation algorithm to determine the mini-model and to compute its answer for the query point. The article describes a learning technique for the MM-method and presents experiment results that show effectiveness of mini-models.
CITATION STYLE
Pietrzykowski, M., & Piegat, A. (2015). Geometric approach in local modeling: Learning of mini-models based on n-dimensional simplex. In Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science) (Vol. 9120, pp. 460–470). Springer Verlag. https://doi.org/10.1007/978-3-319-19369-4_41
Mendeley helps you to discover research relevant for your work.