確認無線網路基地台發射端問題,通常經由管理者親自前往該地區進行測試,或是由使用者主動通報問題;此種無線區域網路管理模式較不具效率。倒傳遞類神經網路具有好的分類能力,已廣泛的應用於解決不同領域的分類問題,但尚未有無線區域網路管理問題之相關應用﹔因此,本研究應用主成份分析、判別分析和Logistic迴歸分析對無線區域網路管理後端的網路安全事件記錄(Security Issues in Network Event Logging, SYSLOG)進行重要屬性挑選,透過倒傳遞類神經網路進行分析,以提供給管理者判斷遠端使用者與無線網路基地台的連線狀態之相關資訊。模擬結果顯示,應用主成份分析挑選重要屬性與倒傳遞類神經網路於無線區域網路之連線狀態分析,為一個可行且可提高無線網路管理效率之方法。
To Confirm the Radio problem of Access Point, it usually goes to far away to test by managers personally or notify the problem by users. It is not an efficient mode of the wireless network management. Back-propagation neural network has good performance on classification, and has been used extensively in many applications. However, There are no relevant application of the problem on the wireless network management yet. So, this research uses Principal Components Analysis、Discriminate Analysis and Logistic Regression Analysis to select Key factors from SYSLOG(Security Issues in Network Event Logging), then uses Back-propagation neural network to analysis Key factors to provide the connect status between users and Access Points to wireless network manager. According to the results, it based on the Principal Components Analysis to select Key factors and Back-propagation neural network to analysis connect status as a feasible and can improve the efficiency of wireless network management methods.