企業動態環境中事件序列之遞增式探掘

Incremental Mining of Event Sequences from Dynamic Business Environments

吳欣怡(Shin-Yi Wu) ； 劉瑞瓏(Rey-Long Liu)

《資訊管理學報》 11卷2期 (2004/04) Pp. 35-60

https://doi.org/10.6382/JIM.200404.0035

摘要

對於多數企業管理者而言，企業環境資訊的搜集與分析是相當必要的。適切地分析出管理者關心之事件出現之順序關係，可讓管理者及早預測重要事件的發生，及時做出適當的應對與防範。由於環境多變，遞增式探掘(incremental mining)對企業環境事件序列探掘而言相當重要。然而，由於環境事件資訊之收集往往具有更高之不確定性（例如多來源之資料登錄與收集、網路忙線、與伺服器未正常提供服務等問題），使得每次搜集而得之事件資訊不必然按其發生先後順序排列，造成事件探掘機制原先所探掘之事件序列遭到破壞，影響探掘結果之正確性。本研究就此問題提出一個解決策略，以遞增式的方式探掘處理事件序列遭破壞的問題，不必重新探掘亦能確保探掘結果之正確性。為驗證此機制之實際貢獻與效能，我們在實際網際網路上搜集企業環境資訊，並進行一系列之實驗。實驗與研究結果顯示，本研究所提出之方法能成功地處理事件序列遭破壞的問題，並在效率上而言，也優於傳統非遞增式的探掘機制。

關鍵字

遞增式探掘 ； 序列探掘 ； 企業環境資訊 ； 資料探掘 ； 事件時間順序

並列摘要

Collecting and analyzing the event sequences from business environments is increasingly important for most businesses. Identifying the sequences of critical events may help managers to identify significant implications earlier, and accordingly respond to the implications in a timely manner. Since businesses environments are intrinsically ever changing, incremental mining of sequences is essential. However, due to several reasons (e.g. multiple sources of information collection, network congestion and heavy loading of various information servers), the collected information pieces can not always be in temporal order. This problem breaks the sequences mined before, and thus deteriorates the correctness of mining. In this paper, an incremental mining technique iSMART is developed to tackle the problem. Instead of re-mining the whole database once an out-of-order event is collected, iSMART incrementally examines necessary parts of the database only. To empirically evaluate the performance of iSMART, we conducted experiments on the Internet. Theoretical analyses and experimental results indicate that iSMART may successfully tackle the sequence-breaking problem with good efficiency.

並列關鍵字

incremental mining ； sequences mining ； business environmental information ； data mining ； temporal order of events

參考文獻

Agrawal, R.,Srikant, R.(1995).Mining sequential patterns.Proceedings of the Eleventh International Conference on Data Engineering.(Proceedings of the Eleventh International Conference on Data Engineering).

Basu, S.,Mooney, R. J.,Pasupuleti, K. V.,Ghosh, J.(2001).Evaluating the Novelty of Text-Mined Rules using Lexical Knowledge.the Seventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.(the Seventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining).

Dörre, J.,Gerstl, P.,Seiffert, R.(1999).Text mining: finding nuggets in mountains of textual data.Proceedings of the fifth ACM SIGKDD international conference on Knowledge discovery and data mining.(Proceedings of the fifth ACM SIGKDD international conference on Knowledge discovery and data mining).

Fukuda, F. H.,Passos, E. L. P.,Pacheco, M. A.,Neto, L. B.,Valerio, J.,Roberto, V. De,Antonio, E. R.,Chiganer, L.(2000).Web text mining using a hybrid system.Proceedings of Sixth Brazilian Symposium on Neural Networks.(Proceedings of Sixth Brazilian Symposium on Neural Networks).

Garofalakis, M.,Rastogi, R.,Shim K.(2002).Mining sequential patterns with regular expression constraints.IEEE Transactions on Knowledge and Data Engineering.14(3),530-552.