| Summary: | 碩士 === 國立中興大學 === 資訊科學與工程學系 === 105 === In recent years, due to the booming development of social media and mobile devices, internet audio and video platforms, e.g., Youtube, have become the popular channels for gaining information. We observed that Youtube was ranked No. 3 in the Alexa Internet’s survey of top sites in Taiwan, so more and more programs were broadcast not only in television but also in Youtube. When people search TV programs and videos, the search results are limited by the description of title and introduction. Therefore, it is an important issue to provide more detail semantic description for videos and TV programs. It means that how to make the video data or TV program data machine-understandable? Currently the general web content (e.g., HTML) is convenient to people browse information. Therefore, we should use semantic web technology (e.g., Linked Data) to make machine understand the web contents or the video program contents.
In this thesis, we design the Resource Description Framework(RDF) for transforming TV programs to linked open data. In the proposed RDF for TV programs, we defined the RDF types, properties, data types of the property values, and the description. The transformation of most properties was easy to deal with, so we focus on generating the special properties, such as HasLabel, SeeAlso. We use named entity recognition and keyword extraction to find TV program’s semantic labels and then linked these labels to DBpedia and Wikipedia.
To demonstrate the merits of the proposed system, we use the courses offered by Taiwan School of Formosa Television as demo cases. The results show that our system can provide more detail knowledge to users
|
|---|
