使用資料探勘方法發掘PKI應用的關鍵成功因素及績效指標

Applying Data Mining Approach to Discover the Critical Successful Factors and Key Performance Indicators in PKI Application Deployment

董伃菁(Yu-Ching Tung)

指導教授 ： 邱昭彰

共同指導教授 ： 徐綺憶

元智大學/資訊學院/資訊管理學系/碩士(2006年)

https://doi.org/10.6838/YZU.2006.00044

摘要

近十年來，因著資訊電子化的發展與進步，愈來愈多的企業已奮力朝向以e-business為經營主軸的方向前進。電子化的企業主要具有以下幾點特色：(1) 逐漸以電子文件取代傳統紙本文件；(2) 資訊交換流程由數天縮短到數小時之間；(3) 異業的結合；(4) 即時的資訊更新；(5) 資訊量累積速度增快。在這樣的一個快速且變動的環境下，人與人之間的信任已不能單單靠紙本的契約來維繫，且當交易發生爭議時，所帶來的企業財務或信用損失將是大到難以估計的。因此，以資訊系統為基礎下的企業遂逐漸有PKI的觀念開始產生。所謂PKI，就是一種網路環境下的契約，利用電子化的法律約束（數位簽章及憑證），使網路各端的成員所進行的協議都能夠被即時記錄起來，並使交易之訊息達到高安全性，只有互相交易的對象能夠認可該交易行為，於事後不可否認該交易行為的發生，更重要的是，這樣的架構可以減少駭客的破壞，使個人或企業的權利更受到嚴密的保障。
在公開的網路交易環境下，因雙方交易的訊息隨時有可能被第三者竊聽或竄改，因此便需要一個信任機制來維繫企業間的信任關係，這種信任機制就不再單純只是幾位元的帳號/密碼登入機制，而是需要更複雜的高位元訊息加密/解密演算法，以及更嚴密的法律條約來規範網路交易的權力與賠償等規章。台灣地區的電子化企業發展截至目前為止，已有多家企業之規模及運作十分成熟且穩健，不僅是在同質企業間具有良好的聲譽和模範，在異質企業間，更吸引了大批廠商合作，所以這些企業具有極有利的條件可以發展公開金鑰基礎建設 (Public Key Infrastructure, PKI) 架構，以創造更高的有形及無形收益。
在本研究中，則藉由過去臺灣區的PKI B2B應用案例探討，利用一些於學界已然發展成熟的統計分析方法來分析PKI推行之較具代表性的關鍵成功因素與績效指標，並以資料探勘 (Data Mining, DM) 的預測方法找出一合理的預測模型 (Predictive Model) 來解釋企業於PKI的推行是否能夠成功，試圖提供業界往後欲推行PKI的參考指標及成功模式。

關鍵字

貝氏網路 ； 關鍵成功因素 ； 遺傳演算法 ； 績效指標 ； 粒子群演算法 ； 結構方程模式

並列摘要

Use e-business in order to deal in the environment of information science and technology making fast progress and change of the main shaft in enterprises. Hense, the interpersonal trust can't already only maintained by the agreement of one paper, and when dispute in trade, enterprise's financial affairs or credit loss brought will be great and difficult to estimate to arrive. So, the enterprises which are based on information system gradually have the idea of PKI. This idea about PKI is one of the agreement of the network environment which utilizes electronic restrain of law (digital signature and certificate authority), and make the agreement carried on by members who exist in every site of the network are, and it could be written down immediately, make the transaction be reached to the superior grade of security, therefore, only people who transact with each other could authorize the transactional behavior, and they have to take the responsibility about the behavior afterwards; the more important things are, such structure can reduce the hacker's destruction, make the rights of individual or enterprises receive the tight guarantee even more. The e-businesses in Taiwan have great scale and stable operations so far, and they are not only possessed good reputation among homogeneity enterprises but also attract a large number of enterprises to cooperate with each other among various enterprises, so these enterprises may have beneficial terms to develop PKI, in order to create the higher tangible and invisible benefits.
In this research, we explore the cases about B2B PKI of Taiwan in the past, and utilized various mature statistical analysis method to analyze the representative factors which conclude CSFs and KPIs; therefore, we utilized the predictive method of data mining field to find out rational models to explain that the implementation about PKI of enterprises could be successful or not could be, and we attempt to offer the models, which could be used by the enterprises that attempt to carry out the PKI, which is refer to the incators and successful model in the future.

並列關鍵字

Bayesian Networks (BN) ； Critical Successful Factors (CSFs) ； Genetic Algorithms (GAs) ； Key Performance Indicators (KPIs) ； Particle Swarm Optimization (PSO) ； Structural Equation Modeling (SEM)

參考文獻

[1] 林禎吉，「公開金鑰基礎建設之研究」，國立成功大學電機工程學系博士論文，2003。

[17] Abunawass, A.M., Bukhres, O., Fisher, T.G., & Magel, K., “A First Undergraduate Course in Neural Networks”, ACM SIGCSE Bulletin, Vol. 22, Iss. 1, Feb 1990.

[18] American Bar Association (ABA), “PKI Assessment Guidelines (PAG),” Information Security Committee, Public Draft for Comment v0.30, 2001.

[19] Armstrong, J.S., & Collopy, F., “Error Measures for Generalizing about Forecasting Methods: Empirical Comparisons,” International Journal of Forecasting, pp. 69-80, 1992.

[20] Bagozzi, R.P., & Yi, Y., “On the Evaluation Structural Equation Models,” Academic of Marketing Science, Vol. 16, pp. 74-94, 1988.