透過您的圖書館登入
IP:3.21.233.41
  • 學位論文

自組織映射圖類神經網路於二類電信經銷商與客戶分類之運用

The Application of Self-organizing Map to Clustering Type II Telecommunications Business and Customers

指導教授 : 黃有評
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


基於台灣電信自由化政策,以及因應台灣加入WTO時所做的承諾,交通部電信總局於2001年7月正式開放第二類電信事業經營轉售業務,加入全球ISR業務開放之行列。隨著電信業務開放,市場也應使用者需求,快速增加許多配套方案,使得使用者能依個人需求選擇適合之服務。二類電信業務包含之服務與一類業者類似,但因為轉售業務的特殊性,二類電信業者能夠彈性的反應市場需求,進而產生各類新業務,最常見的就是語音通信時,當上游路由供應商價格變動時,快速反應在使用者身上。一般而言,上游路由供應商對於價格的變動,主要是因為市場競爭,通常之反映就是價格調降,因此使用者能以更加便宜的費用,購買更多服務。相對於一類電信的費率結構,使用者能以較為便宜之價格,向二類電信業者購買相似的服務,市場上一般通稱為”節費”。 本文主要討論如何將二類電信業者的使用記錄(CDR)加以分類、整理,進而找出經銷商(業務人員)與使用者的關係及使用者撥打行為分類,推論出相似之使用者可能之撥打行為,再依推論之資訊,提供給業者作為業務推展時之參考依據。本文使用類神經網路中自組織映射圖模型,將使用者之通話紀錄予以分群,再分析各分群中相關資料,推論使用者與經銷商(業務人員)之間之關係,及使用者行為。

並列摘要


The Directorate General of Telecommunications of Ministry of Transportation and Communications, based on the commitment for joining WTO and under the cooperation with the policy of the telecommunication liberalization of Taiwan, opened the Type II telecommunications business in July 2001. Following the opening of the telecommunication business, the market offers a lot of business models in conformity with the users’ demand, enabling users to choose the suitable service in accordance with the personal demand. Although the services offered by the Type II telecommunication business are similar to the Type I telecommunication business, the Type II telecommunication service providers can quickly respond the market demand. The most common response is to reduce the price while the cost is reduced. Generally speaking, because of the market competition users can buy more services with cheaper price. Relating to the rate structure of Type I business, users can buy the same services with cheaper price, generally called “Saving Money” on the market. This thesis discusses mainly on how to classify users Call Detail Record (CDR) to find out the relationship between users and agents, the behavior of users, similar user’s possible behavior and so on. This thesis applies the self-organizing map model to clustering CDR and to analyzing the CDR.

參考文獻


[2] Interactive Voice Response, http://en.wikipedia.org/wiki/IVR
[8] M.R. Anderberg, Cluster Analysis for Applications, Academic Press, New York, USA, 1973.
[9] Web content mining, http://www.cyberartsweb.org/cpace/ht/lanman/wcm1.htm
[10] S.C. Hui and G. Jha, “Data mining for customer service support,” Information and Management, vol.38, no.1, pp.1-13, October 2000.
[1] Administrative Regulations on Type II Telecommunications Business, http://www.dgt.gov.tw/english/Regulations/dgt42/dgt42-Type2-3609.htm

延伸閱讀