Aiming at the complexity and uncertainty in the interpretation of rocky desertification from multi-source data, this paper promotes fuzzy Bayesian network embedded Gaussian mixture model(GMM) for extracting rocky desertification information.This model make a fuzzy quantization for continuous variable through GMM, use the convex function of multiple Gaussion density functions to fit the "true" distribution of the data better, and avoid variable's discretization in traditional Bayesian network. All nodes's parameter are then integrated utilizing naïve Bayesian network. Experiments indicate that this model have high accuracy than hybrid Bayesian network and with research value in data mining of multi-source data.

Date Published: 15 October 2009
PDF: 8 pages
Proc. SPIE 7492, International Symposium on Spatial Analysis, Spatial-Temporal Data Modeling, and Data Mining, 74921V (15 October 2009); doi: 10.1117/12.838275

Published in SPIE Proceedings Vol. 7492:
International Symposium on Spatial Analysis, Spatial-Temporal Data Modeling, and Data Mining
Yaolin Liu; Xinming Tang, Editor(s)