使用單一預測模型在空間現象的預測上，即便有不錯的整體預測率，但不保證能在個別的像素或網格上產生良好的預測結果。本研究提出一個新的策略，將不同設計哲學的模式，包含決策樹演算法、類神經網路、貝氏分類器以及案例式概念學習等四個模式加以整合，並且以資料倉儲為資料探勘的基礎平臺，各個模式以及水平、垂直整合之預測結果分別被加以評估及比較。透過這樣的整合方式不僅可以減少模式的不確定性，更提升了預測上的正確性。再以集集大地震引致山崩之空間資料庫為案例，來進行資料探勘預測模式效能的評估。獲致良好的結果，相信類似的方法論可以應用在其他具有豐富空間資料的環境議題研究上。

Using a single model to forecast spatial phenomena will not produce good estimation in the prediction of individual pixels, even with good overall accuracy. A new strategy, which combines several models based on different philosophies, not only reduces the uncertainty of predictive modeling but also improves its accuracy. This study integrates a Decision Tree algorithm, the artificial Neural Network, the Bayes Classfier, and Exemplar-based concept Learning, with each individually applied to a spatial data warehouse. The results of each model and two kinds of modeling-integration methods, including horizontal integration and vertical integration, were then evaluated. In a case study, chose Chi-Chi earthquake-induced landslide to test the prediction accuracy and obtained good results. We believe that the same methodology can also been used in other cases of environment issues for which there is plentiful GIS digital data.