以樹狀序列挖掘企業系統效能規則

Mining Enterprise System Monitoring Trees with Sequences

林艷(Yen Lin) ； 韋俊仲(Chun-Chung Wei) ； 鄭明松(Ming-Sung Chen) ； 許秉瑜(Ping-Yu Hsu)

《資訊管理學報》 11卷3期 (2004/07) Pp. 155-178

https://doi.org/10.6382/JIM.200407.0155

摘要

企業系統被視為企業運轉的基石，在企業中扮演著極重要的角色，因此系統的穩定性相當地重要。本文主要是結合資料挖礦技術中的序列關聯規則與有根樹資料結構，發展出適合Tree-Based系統效能資料特性的演算法－SPT(Sequence of Performance Trees)。針對樹狀序列資料結構，以及數值屬性資料的處理技巧提出的SPT演算法，可針對系統管理者感興趣的階層進行挖掘，找出的關聯規則可以幫助系統管理者從大量的系統效能資料中找出潛藏的訊息，同時可以發現所觀測現象發生的序列變化，這些資訊可以提供系統管理者調校系統、分析和矯正錯誤、增進系統效能，讓企業系統運作更加穩定及可靠。
本文中SPT演算法應用在企業系統的效能分析上，可以幫助系統管理員整理出很多有意義的特徵規則或系統效能的型態(patterns)。

關鍵字

資料挖礦 ； 樹狀序列資料 ； 關聯規則 ； 企業系統效能型態

並列摘要

For enterprises implementing enterprise systems, the systems become the main pipeline where information flows integrated. If the pipelines get clogged, the whole enterprises can grind to a stall. Therefore, maintaining the stability and availability of enterprise systems become one of the most important duties of MIS department. To properly maintain the systems, not only do system administrators need to react to problems but also need to predict the patterns of system performance and act proactively.
The paper proposes to find the patterns with SPT, Sequences of Performance Trees. The patterns and underlying data are represented as sequences of monitoring trees. A tree is a collection of system performance indicators in a particular moment. Trees are organized to sequences to keep system performance in a consecutive time intervals. Since a system administrator may not be interested in the evolution patterns of entire trees, the paper also proposes an interest vector to pick interested subtrees and perform association rule mining on those subtrees. With the addition of the interest vector, the definitions of support and sequence equivalence are all modified. A demo system is also shown in the paper.

並列關鍵字

Data Mining ； Tree-based structure ； association rules ； enterprise system ； Performance Pattern

參考文獻

陳仕昇、陳彥良、許秉瑜(2001)。在序列式資料中挖掘序列規則。資訊管理學報。6(2)

陳仕昇、許秉瑜、陳彥良(1997)。以可重覆序列挖掘網路瀏覽規則之研究。資管評論。

陳智宗、陳振明、許秉瑜(2000)。以資料挖礦法挖掘多屬性序列式資料規則之研究(碩士論文)。中大資管所。

盧靜婷、陳彥良(2000)。在DAG中挖掘家族特徵規則(碩士論文)。中大資管所。

Agrawal, R.,Imielinski, T.,Swami, A.N.(1993).Mining association rules between sets of items in large database.Proceedings of the 1993 ACM SIGMOD International Conference on Management of Data.(Proceedings of the 1993 ACM SIGMOD International Conference on Management of Data).: