資訊管理學報
黃國禎;林宗良;
頁: 171-193
日期: 2005/01
摘要: 隨著網路普及率的提昇及上網人口年齡的下降，世界各國對網站內容分級的重視程度也隨著增加。在推網站內容分級的過程中，除了訂定分級標準之外，最重要的研究主題之一，即是分級伺服器的建置與維護策略。由於分級資訊變化快速，而眾多網路用戶在瀏覽網站內容時又必須隨時查詢分級資訊；因此，如何在大量資料存取的需求下，提供高效率的網站內容分級資訊，將是影響成敗的重要因素之一。過去以單一伺服器建立的分級資訊管理方式，在大量資料存取時，將造成網路的擁塞及長時問的等待；因此，本論文中，我們運用基因演算法來解決分散式分級資訊的規劃及配置問題，以達到提供大量且快速分級資訊服務的目的；同時，並實際建置一套分散式分級伺服器，與原有的proxy系統進行整合，以提供分級服務。由大量的資料存取測試的結果發現，本論所文提出的分級資訊管理方式，可以讓網路用戶在不需改變使用習慣的前提，即可獲得有效率的分級服務。
關鍵字: 網路內容分級;PICS;資訊管理;基因演算法;
Abstract: As computer network becomes popular and the average age of the Internet users decreases, website labeling has become an important and challenging issue. In addition to the development of website content rating standard, it is very important to study the maintenance and management of the website labeling information. As the contents of a website may change frequently, and a tremendous number of users need to obtain the information when browsing web pages, the policy of the content-labeling information management will significantly affect the efficiency of the whole website labeling system. In can be seen that installing the entire website labeling data to a single server will lead to network traffic jam and server overloading. To cope with these problems, in this paper, we propose a distributed model for managing website labeling information. To achieve maximum throughput and optimal load balance for the distributed content rating information, we develop a genetic algorithm to allocate website labeling information for each server. Some experiments on a large amount of test data have shown that the genetic approach can efficiently allocate website labeling information and provide desirable results.
Keywords: web content labeling;PICS;information management;genetic algorithms;
瀏覽次數: 10310 下載次數: 80
引用 導入Endnote
頁: 171-193
日期: 2005/01
摘要: 隨著網路普及率的提昇及上網人口年齡的下降，世界各國對網站內容分級的重視程度也隨著增加。在推網站內容分級的過程中，除了訂定分級標準之外，最重要的研究主題之一，即是分級伺服器的建置與維護策略。由於分級資訊變化快速，而眾多網路用戶在瀏覽網站內容時又必須隨時查詢分級資訊；因此，如何在大量資料存取的需求下，提供高效率的網站內容分級資訊，將是影響成敗的重要因素之一。過去以單一伺服器建立的分級資訊管理方式，在大量資料存取時，將造成網路的擁塞及長時問的等待；因此，本論文中，我們運用基因演算法來解決分散式分級資訊的規劃及配置問題，以達到提供大量且快速分級資訊服務的目的；同時，並實際建置一套分散式分級伺服器，與原有的proxy系統進行整合，以提供分級服務。由大量的資料存取測試的結果發現，本論所文提出的分級資訊管理方式，可以讓網路用戶在不需改變使用習慣的前提，即可獲得有效率的分級服務。
關鍵字: 網路內容分級;PICS;資訊管理;基因演算法;
A Genetic Approach to the Management of Website Content Rating Information
Abstract: As computer network becomes popular and the average age of the Internet users decreases, website labeling has become an important and challenging issue. In addition to the development of website content rating standard, it is very important to study the maintenance and management of the website labeling information. As the contents of a website may change frequently, and a tremendous number of users need to obtain the information when browsing web pages, the policy of the content-labeling information management will significantly affect the efficiency of the whole website labeling system. In can be seen that installing the entire website labeling data to a single server will lead to network traffic jam and server overloading. To cope with these problems, in this paper, we propose a distributed model for managing website labeling information. To achieve maximum throughput and optimal load balance for the distributed content rating information, we develop a genetic algorithm to allocate website labeling information for each server. Some experiments on a large amount of test data have shown that the genetic approach can efficiently allocate website labeling information and provide desirable results.
Keywords: web content labeling;PICS;information management;genetic algorithms;
瀏覽次數: 10310 下載次數: 80
引用 導入Endnote