Scalable parallel algorithms for surface fitting and data mining

This paper presents scalable parallel algorithms for high-dimensional surfa ce fitting and predictive modelling which are used in data mining applicati ons. These algorithms are based on techniques like finite elements, thin pl ate splines, wavelets and additive models. They all consist of two steps: F irst, data is read from secondary storage and a linear system is assembled. Secondly, the linear system is solved. The assembly can be done with almos t no communication and the size of the linear system is independent of the data size. Thus the presented algorithms are both scalable with the data si ze and the number of processors. (C) 2001 Elsevier Science B.V. All rights reserved.