DTV Personalized Information Filtering System Design

With the popularity of DTV and maturity of the related digital technology, apart from the traditional TV service, a variety of completely new value-added services on the digital broadcasting TV net have derived and users can enjoy omni-bearing information services, such as digital TV, broadcasting data, Internet connection, etc.The thesis aims at constructing DTV Text Stream Filter System (TFS) on the terminal platform, which offers the users an efficient means to receive digital information. The functions of TFS are to filtrate and store up text information submitted by DTV according to the user's interests. In technology it is similar to Personalized Information Filting, while the difference lies in the fact that it is a home, tidy and small text information filter which works with Set-top Box as the working platform and with text information of TS as operand. It is applicable in the transition period when broadcasting TV network is remolded from single-direction to two-direction. Even after DTV realizes two-direction, TFS is an important technique that improves the efficiency of DTV value-added service.The thesis has done the following work:Firstly, it has done deep research on the standards of MPEG-2, DVB and data broadcasting that have close relation with the study, and made selective analysis of TS stream's structure and the method to extract text information from TS stream.Secondly, it analyzes the hardware's and software's structure of TFS's working platform, the Set-top Box.Thirdly, it puts forward the way to judge the degree of correlation between paragraphs in a text by using "Concept Field" attribute of key words in the dictionary, which can obtain the balance between filtration accuracy and screening efficiency to some degree.Fourthly, it proposes, in consideration of the features of present broadcasting TV network and the users' personalized information demand, a TFS design and its realization method that endows TFS with the function of automatic analysis of the users' interest tendency.Finally, it prospects the further research of the study.