地震防救災文獻案例式查詢系統

A CASE-BASED LIBRARY QUERY SYSTEM FOR EARTHQUAKE HAZARDS PREVENTION

台灣是一個地震災害頻仍的地區，為了防救災的需要，長久以來，研究單位也累積了相當多的成果。然而，當吾人遇到一個新的問題情境時，如何從汗牛充棟的研究報告中，快速而正確地找尋到適宜的研究成果，以供參考，便成了一個相當重要的課題。本研究利用「案例式系統」的技術來嘗試解決這個問題。本研究將地震相關的研究報告歸納成13種屬性，並且透過專家訪談的方式，選定查詢的關鍵指標及一般指標，同時以Caspian軟體進行實例測試其可行性。其結果顯示，當問題情境定義清楚時，只用一個關鍵指標即可以找到相當有用的資料。

Taiwan suffers a lot of crises from earthquakes. To prevent hazards, there are many researches that have been done for a long time. However, it is a critical issue that retrieving an appropriate research paper from a huge amount of reports. This research employs the technique of case-based reasoning, where each case corresponds to one research report, to tackle the difficulty. Furthermore, we use 13 attributes for these research reports, and identify critical and ordinary indices by interviewing experts. Free software, Caspian, is also used to implement a prototype for testing the appropriateness of attributes, weights, and indices in the case-based system. It is found that single critical index is sufficient for a well-defined scenario.