Study On Authority Control Of Social Tagging And Its Application

Social tags are similar to keywords or metadata used in the traditional information resources organization. Born in the network environment, social tags are created by public users, and they contain a wealth of semantic information, so applying social tags to text automatic classification has some practical significance. But due to the excessive freedom of the usage of social tags, the quality of social tags is relatively lower, which leads to a large number of redundant tags, synonyms tags, and so on. All these problems will affect the efficiency of the social tags’ text classification. So, this paper aims to propose a "core controlled, shell uncontrolled" text classification model on the basis of Chinese books’ social tags’feature analysis, which means to use keywords to control the quality of social tags through the establishment of "tags-keywords" concept space. At last, this paper verifies the rationality and feasibility of this method through some specific experiments.The specific contents of this paper are as follows:(1) We summarized the status of the research on social tagging and its related knowledge and discussed the research progress of the authority control of social tags, as well as the specific applications of social tags.(2)We made an analysis of Chinese books’social tags, and through the indicators of length distribution, usage distribution, frequency distribution, scale, average length, and class average depth of indexing of these tags, we proved the similarity of social tags and keywords. So it is possible to apply the method dealt with keywords to the process of social tags. We also put forward the way to control the quality of social tags through the establishment of "tags-keywords" concept space.(3)According to the features of social tags, we introduced the concept of concept space, and established a "tags-keywords" concept space, so as to fulfill the object of using keywords to control the quality of social tags.(4)As for the specific application of social tags, we particularly applied social tags to Chinese books’automatic classification, and proposed a "kernel controlled, shell uncontrolled" classification model according to the principle of linguistic intelligence, which pointed to improve the automatic classification’ efficiency though some control methods done at the background without affecting the users’ normal operation at the front.The innovations of this paper have the following two aspects:(1)Through the investigation and analysis of the characteristics of Chinese books’ social tags, we confirmed that social tags have some same characteristics of natural languages, so it provided a theoretical basis for natural language processing methods to be applied to the process of social tags, thereby we introduced the way of concept space to build "tags-keywords" concept space of Chinese books to achieve social tags’ authority control.(2)This paper presented a "kernel controlled, shell uncontrolled" classification model for Chinese books, and fulfilled the goal of controlling the quality of social tags in the kernel of the social tagging system, and keeping the users’ habits outside the system.