現今網際服務（Web Service）已經成為許多大型企業用來整合商業流程的主 要技術，如何從現有的服務中發現複合服務的議題成為熱門的研究領域。本篇論 文主要是利用資料探勘領域中兩種方法來探勘網際服務使用者記錄，以分析網際 服務之間的關係，第一種方法，多層關聯規則探勘方法（Multilevel Association Rules Mining）可探勘出常用的服務組合，且可發現服務間較高層次上的關係； 第二種方法，序列模式探勘（Sequential Pattern Mining）可探勘出常用序列的服 務組合關係。前者產出結果為無順序性的相關服務組合，可作為一種建議，讓使 用者自行選擇使用；後者產出結果為具有順序性相關的服務組合，並有助於整合 成真實的商業流程。實驗顯示我們提出的方法具有實用性、彈性及效率，根據本 方法探勘的結果促使容易整合複合服務組合。 Web Service Technology is being applied to organizing business process in many large-scale enterprises. Discovery of Composite Service, therefore, has become an active research area. In this paper, we propose two methodologies in data mining area to analyze the relationship among these web services from web service usage log. First, Multilevel Association Rules Mining is used for discovery of frequently used sets of web services. Additionally, it can extract high-level relationships among web services. Second, Sequential Pattern Mining is used for discovery of the sequence of web services. The former produces unordered sets of web services which can be used as suggestions to the user. The latter generates time-ordered sets of web services which can be exploited to integrate into a real business process. The empirical result shows the proposed methodologies are useful, flexible, and efficient. It is able to integrate simple web services into a composite service according to the mining result of the proposed approach.
