Dissertation > Excellent graduate degree dissertation topics show

Research on Media Catchwords Extraction

Author: WuBaoZhen
Tutor: HeZuoZuo
School: Central China Normal University
Course: Applied Computer Technology
Keywords: Buzzwords Media buzzwords Annual media buzzwords Term Extraction
CLC: H109.4
Type: Master's thesis
Year: 2009
Downloads: 292
Quote: 2
Read: Download Dissertation

Abstract

Media buzzwords quickly popular, widely spread the word in the newspapers, television, radio, Internet and other media. The annual media buzzwords can truly reflect high-level overview of the year's domestic policy initiatives, social events, people focus as well as changes in the international situation. The different annual media buzzwords changes can reflect the changing social dynamics and mass psychology. The access media buzzwords research is an important part of the the language dynamic monitoring work has important academic value and a broad and deep social and cultural significance. This paper conducted in-depth study of media buzzwords acquisition method, the main work includes the following aspects: first, the media buzzwords characteristics analysis. Analysis of popular state media buzzwords, media buzzwords popular feature model. Second, to propose a scoring model automatically determine the media buzzwords. The introduction of the words commonly used degrees attributes and time attributes, and change attributes, and thus quantify these attributes to get media buzzwords score formula. Automatically in accordance with the rated values ??of the candidate words and lasted trend curve to get the media buzzwords. Third, the design and to achieve annual media buzzwords acquisition system. The system uses Web crawler to download the news pages of the 2007 five portals (Netease, Sohu, Sina, Tencent, Tom) (10,642 MB) is a research corpus, and their pretreatment text segmentation and segmentation All word string filter, at last, the paper proposes the use of buzzwords scoring model screened annual media buzzwords. Fourth, in the process of obtaining the set of candidate words, the proposed combination method of obtaining a statistical and rules. First full segmentation-based approach to obtain all possible word string, and then use to turn all word string filter candidate popular words set the filtering rules based on linguistic knowledge, the garbage string filtering rules and heavy filtering rules based on TF · IDF rights. The annual media buzzwords acquisition system, method and design of this study has applied to the National Language Resource Monitoring and Research Center, online media language sub-centers the 2007 media buzzwords extract and released. And achieved good results, which verify the of media buzzwords characteristics model and automatic acquisition system validity. The system can be released for media buzzwords to provide objective, real and high-quality data, save a lot of manpower and resources.

Research on Media Catchwords Extraction

Abstract

Related Dissertations