建構完整的漁業資訊分享熱線:三個異質性資料庫的整合經驗

Abstract

漁業資訊分享熱線為一個以網路為基底的資料倉儲建置過程，主要是被設計來整合三個異質性的關連性資料庫。漁業資訊分享熱線中的資料流通在本質上可以區分為三個階段。在整備階段時，透過檔案轉移服務以及資料轉換服務的機制將漁業資料傳送到資料倉儲中。在前段階段，使用者可以透過線上分析處理的介面來擷取資料倉儲中的資料。除此之外，在資料探勘階段可以透過共整合分析技術擷取時間序列資料來探究其統計時變的性質。在進行整合之前，從最初的資料收集到最後的共整合分析必須花費數週的處理時間，但是在建立漁業資訊分享熱限制之後，則僅需花費數小時的時間就可完成相同的任務。採用網路為基底的資料倉儲不僅提供一個整合性與一致性的資料來源，並且提供一個高效率模式來處理與分析漁業資訊。

A web-based data warehouse entitled Fishery Information Sharing Hotline (FISH) has been designed and implemented for integrating three heterogeneous relational databases. Essentially the data flow in FISH is divided into three stages. At the backend stage, fishery data are sent and transformed into the data warehouse though File Transfer Service (FTS) and Data Transformation Service (DTS). At the front stage, a user can retrieve the consistent data from the data warehouse through an On-Line Analytical Processing (OLAP) interface. Moreover, the retrieved time series data are analyzed for exploring their non-stationary properties using co-integration technology at the data-mining stage. Before this integration, the processing time usually takes several weeks from the initial data collection to the final co-integration analysis. After the establishment of FISH, it only takes a few hours for achieving the same task. The web-based data warehouse approach provides not only an integrated and consistent data source but also a highly efficient way for processing and analyzing fishery data.