模糊專家系統在新聞編排推薦之應用－以三立電視新聞台為例

電視新聞部門每天提供收視觀眾多樣化而且充足的重要訊息，新聞編輯人員面對龐大的新聞資訊內容，必須根據新聞的重要性，試著去分類新聞。該如何決定新聞編排的優先程度？有沒有一個依循的準則可以參考？這是值得研究的議題。傳統上，新聞編排都是由編輯人員依個人經驗或主觀的想法來編排，這樣的作業模式有時候容易出現意見相左的情況。有別於傳統方法，本研究以模糊理論為基礎，建立一套新聞自動編輯系統，這套自動系統由模糊規則庫、模糊推論引擎及使用者介面所構成，藉由本系統，編輯人員可以有效率的編排新聞，並根據新聞的重要性排定播報順序。本研究訪談五位領域專家，並根據他們的專業知識，建立模糊規則庫。最後並提出個案研究及討論實驗結果。個案研究的資料來源為三立電視台九十二年十一月五日當天的一百三十則新聞。

關鍵字

模糊理論；模糊專家系統；新聞編輯

並列摘要

The TV news department offers the audience various and plenty of information daily. News editors face a large amount of news everyday and try to classify the news according to the importance of news. What kind of news is the most important and interesting to the audience? What are the rules to determine the news's priority for reporiting? These issues are interesting research topics. Traditionally, the news editors determine the news priority by their personal viewpoints and experiences intuitively. It may sometimes lead to a conflict among different viewpoints from different editors. Unlike the traditional method, this study proposes an automatic news editing system based on fuzzy theory. The proposed fuzzy expert system consists of a fuzzy rule base, a fuzzy inference engine and a user interface. Using this system, the news editors can efficiently edit subjective-oriented news and therefore arrange the schedule for news reporting. The fuzzy rule base was built based on the knowledge provided by five domain experts. The study also presents a case study and discusses the experimental results. There are 130 pieces of news used in the case study, which were taken from the SETV on November 5(superscript th), 2003.