Geo-Spatial Data Mining and Visualization Based on Texts

• 1、School of Resources and Earth Science, CUMT

Abstract：Recently, Web Text Mining is more and more popular to different kinds of industries, while geographical Data Mining and Visualization based on text belong to an emerging field. After reviewing the status quo of text mining research, the thesis brings forward an application model of geographical Data Mining based on traditional texts, and implements visualization of geographical information extracted from text data mining and inquiry of related attributes of historic scene places, by using Arcview3.2 as modeling platform. This thesis selects the second road in the north of Lushan as an application example research of Text Data Mining Model. First of all, traditional textual contents related to the second road in the north of Lushan are converted to electrical documents and necessary preprocessing is performed. Then, a data dictionary is set up covering the primary scene places and their spatial connection and a code system is established in terms of the geometry classification, namely as points or lines, of the scenes in the case study. Then sketches are drawn and architecture of spatial database is designed. After filling out database tables, both the spatial and attributes data are stored as different themes in Arcview3.2. The experimental model system possesses a series functions, like classification, statistics, and presentation of historic scene places. Moreover, it can be used to make comparative analysis between historic and current scene places. Finally the thesis analyses the weakness of this model and that proposes suggestions for improvement.

Keywords： The second road in the north of Lushan Text Mining Data dictionary Query