| Summary: | 碩士 === 輔仁大學 === 資訊管理學系 === 100 === In order to create service innovation for Web 2.0 video-sharing platforms on the Internet, this research is to recommend the further relative tags and videos utilizes through a large number of user-generated tags and video. Thus, this research firstly adopts ‘Voting-Promotion’ to assess co-occurrence of tags to further determine the recommended list of candidate tags. Furthermore, the tag aggregation methodology includes voting and Borda Count to analyze the candidate tags at the second stage.
According to the tags and videos provided by the Internet users, the co-occurrence analysis can be assessed by query tag and the other tags from the tag’s database to measure the degree of correlation which can analyze the most relevant other tags. The second module, Voting and Borda Count, can conduct the aggregation function to acquire the further candidate tags. The third module is promotion methodology that can rank and rate these recommended videos based on descriptive promotion, stable promotion, and rank promotion measurements. The objective of this research is to assist the users who hope to tag and upload videos to determine choice of tagging. In other words, the user-generated content can be analyzed to recommend for the query uses. Thus, this proposed service system can provide the recommendation information for tagging through the automation of decision making process of tagging at the shared video platform.
However, the Voting-Promotion function can enhance the determination of tagging, it can acquire the tags’ relationships and further predict the degree of exposure when the users tag and upload videos According to system experiments, we use simulated data to analyze the effects of this tagging recommendation model.
In the first experiment, it demonstrates the significant difference between the text retrieval and this vote-promotion methodology. That’s reason why this research adopts the vote-promotion methodology to develop the tagging recommendation service. The second experiment can verify the effects of ranking tags and videos. The vote-promotion can provide the different tagging ranking that derived from the collaborative decision making process. The third experiment shows the significant difference of relative, satisfactory, and helpful among the text retrieval and the proposed vote-promotion methodology. Thus, the experimental results show that Voting-Promotion function was all significant differences to the Text-retrieval function, only the relative of recommended videos is not highly significant differences. From the perspective of service innovation, this research demonstrates how to enhance the exchange value behavior between the sharing user, and propose an innovate service system design that can be systematic and quantify value co-creation of the shared tags and videos.
|
|---|
