Integration of Geo-processing techniques to SmartBase Viewer

This work is intended to integrate geo-processing techniques to the visualization of data mining results, serving as a tool to aid decision-making. As validation, we used a software called SmartBaseG to accomplish data mining and the tool Smartbase Viewer to present the results.

The SmartBase Viewer presents the mining results under varied models, such as the AssociationModel, ClusteringModel, TextModel, and TreeModel. We will concentrate on the tree model. It shows in an objective way all of the concepts generated by the algorithm. However, in some cases, we noticed the need to relate the correlations found by the Data Mininig process with geographic information.

Geo-processing is an alternative to this problem. It can be used in a data mining system when the base to be processed has geographical attributes (latitude and longitude). Thus, by presenting the knowledge generated as being points on a map, we are placing data mining close to the daily of the user. These two approaches complete each other and they present the result in a much more complete manner. What is not noticed by the user in the decision tree, may be visible when we place the information in the cartographic base. And when some component of the result is not contemplated by the map, the concept may be present in the tree.

Thus, we are increasing the quality of the information that will aid at the time of the decision.

The idea is to show the tree and the map concomitantly and for each branch of the tree, that represents a concept, we can visualize the crimes covered by this concept. In it, we have the points of two different branches being drawn on the maps. We have the conclusions generated by the mining algorithm placed on the map and when we observe the characteristics of the area we may reach new conclusions, more complete, on the content of the information present in the database.