Affiliations: Department of Computer Science and Engineering,
University of Nebraska - Lincoln, Lincoln, NE 68588-0115, USA
Abstract: In this paper, we investigate interpolation methods that are
suitable for discovering spatiotemporal association rules for unsampled sites
with a focus on drought risk management problem. For drought risk management,
raw weather data is collected, converted to various indices, and then mined for
association rules. To generate association rules for the unsampled sites,
interpolation methods can be applied at any stage of this data mining process.
We develop and integrate three interpolation models into our association rule
mining algorithm. We call them pre-order, in-order and post-order interpolation
models. The performance of these three models is experimentally evaluated
comparing the interpolated association rules with the rules discovered from
actual raw data based on two metrics, precision and recall. Our experiments
show that the post-order interpolation model provides the highest precision
among the three models, and the Kriging method in the pre-order interpolation
model presents the highest recall.
Keywords: data mining, association rules, interpolation, cross-validation, geo-spatial decision support system
