19th International CODATA Conference
Category: Knowledge Discovery
Monica Wachowicz (Monica.Wachowicz@wur.nl)
Wageningen UR, Centre for Geo-Information, The Netherlands
http://www.geo-informatie.nl/
Traditional exploratory spatial analysis tools are inadequate for handling the
increasing volumes of data and the complexity associated with discovering spatio-temporal
patterns within a Geographic Knowledge Discovery (GKD) process. Some attempts
to develop methods and associated tools for a GKD process have already illustrated
how difficult is to make use of appropriate interaction forms, visual representations
and mining tasks in order to allow users to dynamically construct geographic
knowledge. The major challenge here is to empirically demonstrate that a combination
of geovisualisation and data mining methods can improve
a GKD process and to build geocollaborative environments that take advantage of the best
that each has to offer.
This paper describes the design of a collaborative environment for manipulating spatio-temporal data to find, relate, and interpret interesting patterns in very large land use data sets. We explain how different geovisualisation and data mining methods and functionalities can be combined in order to design a GKD process for the identification and interpretation of the space-time variability in both composition and structure of changing and use patterns. Therefore, a GKD process is proposed in this paper for creating a dynamic method for finding, relating, and interpreting interesting, meaningful, and unanticipated patterns in large land use data sets. The goal is to develop a conceptualisation of a GKD process that involves different users achieving insight about patterns of land use change within a collaborative environment, which facilitates the understanding of these patterns and their relation with the corresponding process in the real world (e.g. urbanisation, deforestation).
A tool prototype (Land Use Change Explorer) has been developed to allow different users to discover land use change patterns and understand their corresponding process in the real world. The Land Use Change Explorer was implemented as a multi-agent collaborative environment. In this environment agents help users to perform a series of steps of a GKD process, namely: selecting land use data sets; selecting proper representations to visualise land use change; choosing sequences of mining operations; geovisualisation functionalities. Use of the prototype is demonstrated for two types of users. They represent two stakeholders: an agricultural policy maker and an urban planner. The aim is to show how their pooled expertise affects the GKD process within a collaborative environment. At this stage, the Land Use Change Explorer prototype supports the abductive mode of reasoning.
Keywords: geographic knowledge discovery, spatio-temporal data mining, geocollaborative environments, multi-agents system, land use change explorer.