運用案例式推理機制建構電腦問題診斷專家系統 -以個案公司為例

Using case-based reasoning mechanisms to construct a computer problem diagnosis expert system–A case study of company

劉正廉(Cheng-lien Liu)

指導教授 ： 李永銘

國立交通大學/管理學院/管理學院資訊管理學程/碩士(2014年)

https://doi.org/10.6842/NCTU.2014.00372

摘要

公司的網管人員除了每天例性行檢查系統是否出現異常之外，還得負責排除公司同仁的電腦與網路之間出現的各種問題。因此建置一套系統管理工具，並且在最短的時間內處理這些問題是首當之急的工作。
本研究是以案例式推理為基礎，利用分群方法並結合文章向量、權重與相似度進行群集分析工作，並且使用貝氏分類法進行分類模型訓練與測試並計算出準確率及召回率值。將歷年來系統解決問題的方法累積記錄成知識庫，配合事件處理的流程制度來有效控管電腦的問題，並且做為日後問題處理的參考依據。
本研究驗證方法以準確率及召回率做為評估的指標，實驗結果顯示準確率及召回率已達95%以上，已達到某種程度的可靠性。藉此期許對企業內部資訊系統於處理電腦異常問題有所幫助，未來更朝向更多元的解決方案供系統管理人員使用，提升整體競爭力。

關鍵字

專家系統 ； 案例式推理 ； 準確率與召回率

並列摘要

In addition to the company's network management in checking whether the system faces abnormal occurences, network administors are responsible for solving computer and network problems. So building a system management tool to deal with these issues in the shortest possible time is the first priproty and urgency of the work. Based on case-based reasoning, this research uses clustering methods combined with article vector similarity weights for cnducting cluster analysis and utlizes Bayesian classification model to train and test dataset and calculate the accuracy and recall value. The system records a cumulative knowledge base according to the problem occuring the the past years and uses event processing system to effectively control the flow of computer problems, which are addressed as references for the future.
In this study, the recall rate and authentication methods are used as indicators to assess the accuracy of the experimental results. It is shown that the recall rate and the accuracy rate are more than 95 percentage which has reached a certain degree of reliability. It is expexted that the system will help enterprise information systems to handle unusual problems. In the future, more diverse solutions can be developed for system administrators to enhance the overall competitiveness.

並列關鍵字

Expert Systems ； Case Based Reasoning ； Precision and recall rate

參考文獻

Zhang, Y.-h., et al. (2011). Optimization of knowledge work based on experts' implicit knowledge mining. Industrial Engineering and Engineering Management (IE&EM), 2011 IEEE 18Th International Conference on, IEEE.

Liao, S.-H. (2005). "Expert system methodologies and applications—a decade review from 1995 to 2004." Expert systems with applications 28(1): 93-103.

Joshi, S. and B. Nigam (2011). Categorizing the Document Using Multi Class Classification in Data Mining. Computational Intelligence and Communication Networks (CICN), 2011 International Conference on, IEEE.

Piao, X., et al. (2010). Research on mining positive and negative association rules based on dual confidence. Internet Computing for Science and Engineering (ICICSE), 2010 Fifth International Conference on, IEEE.

Efrain Turban, D. L., Ephrain Mclean, James Wetherbe (2008). Information Technology for Management 6E.