A Kriging Fuzzy Integral

This paper presents a new mathematical model for integrating information from multiple sources with respect to fuzzy measures. The proposed methodology is based on the optimality principle of an estimation technique of geostatistics known as block kriging. The feature of this analytical solution method for information fusion problems is that it provides an operation on aggregating the available data in such a way as to minimize the expected squared error of estimation, and the estimation error is also computed. The results obtained from this method will be compared with the ordinary weighted mean, Sugeno integral, and Choquet integral. © Elsevier Science Inc. 1997.