<< Click to Display Table of Contents >> Navigation:  Using GeNIe > Geo-processing > Introduction

A growing number of users apply Bayesian networks to processing geographical data. In this case, a typical processing occurs at the level of individual cells of a raster representing a map. Processing amounts to collecting input parameters (used as evidence in a Bayesian network model) from individual maps and producing maps that contain information derived by means of the Bayesian network model. This happens at each cell of the raster map. For example, given the height above the sea level of a map point, average amount of rainfall, average temperature, etc., a Bayesian network may derive the probability of vegetation in that particular raster cell.

It is possible to perform these calculations using SMILE without the geo-processing extensions. In this case, one would have to read data for every raster cell from every input map, process that information using a Bayesian network model, and then produce values for the output cells. Direct support for this calculation makes processing much faster, as there is no overhead required for accessing SMILE for every single raster cell.

This section describes GeNIe's map processing capability.