YouTube音樂影片排行服務之設計與實作

隨著網路科技的進步，資訊的分享形式已從靜態的圖文提升至動態的影音模式，目前流量第一的影片分享網站－YouTube，每分鐘就有120個小時的影片量被上傳，如何在這麼繁多的影片資料中提供使用者當下的熱門影片，讓使用者可以透過影音資訊快速地掌握社會的脈動是一個重要的課題。本論文實作了一個以音樂類別為主的影片排行服務。透過取得類別的詞彙名單後，蒐集與詞彙相關的YouTube影片資訊，運用影片的觀看次數統計出各期間的成長量。依據每日/每週/每月成長量以及觀看次數的排序，得到各類別的排行榜。另外，本服務利用斷詞法與觀察影音內容後所訂出的取詞規則，提供網站管理者一個針對音樂類別產生詞彙的模式。

關鍵字

詞彙擷取；排行服務

並列摘要

As the Internet technology improved, the way people share their life has changed - from the normal pictures/texts to dynamic videos. YouTube, the most popular video-sharing website, has at least 120 hours of videos uploaded every minute. How to know which videos are most popular in such a vast amount of data is an important issue. In this thesis, I designed and implemented a web service to provide a classified ranking service for YouTube musical videos. The feature set of a category is formed by extracting the representative keywords - the names of the singers, from such category. The keywords are then used to crawl the video information for videos in such category. The ranking of the videos are decided by the amount of viewcount increase in a daily/weekly/monthly basis. In addition, the system provides a simple method for the administrator to build the key term list for a specific category.