CityU Institutional Repository >
3_CityU Electronic Theses and Dissertations >
ETD - Dept. of Computer Science >
CS - Master of Philosophy >
Please use this identifier to cite or link to this item:
|Title: ||Automatic tag recommendation for the Web2.0 blogosphere|
|Other Titles: ||Web2.0 bo ke kong jian de zi dong biao qian jian yi xi tong|
|Authors: ||Lee, On Kee (李安基)|
|Department: ||Department of Computer Science|
|Degree: ||Master of Philosophy|
|Issue Date: ||2008|
|Publisher: ||City University of Hong Kong|
|Subjects: ||Blogs -- Design.|
|Notes: ||CityU Call Number: TK5105.8884 .L44 2008|
105 leaves : ill. 30 cm.
Thesis (M.Phil.)--City University of Hong Kong, 2008.
Includes bibliographical references (leaves 101-104)
|Abstract: ||This research proposes a novel approach to automatic tag recommendation for weblogs/blogs. Tagging is used mainly to assist search. However, tagging is a manual process. Different people may use different tags for similar content. To search, users have to make several “intelligent guesses” before they can find interesting blogs. Therefore, it is very useful if we know which tags are most appropriate and popular to use. Web 2.0 was very new when project started; we decided to explore two key Web 2.0 technologies: blogs and tags. In the research, two different approaches for tag recommendation have been proposed. The first approach makes use of collective intelligence extracted from Web 2.0, collaborative tagging as well as word semantics to learn how to predict the best set of tags to use, using a hybrid artificial neural network (ANN). The second approach makes use of Vector Space Model (VSM), similar documents from the web and some statistical methods for tags recommendation. The research object is to design and implement software to automatically suggest tags for a blog entry. The reasons why two algorithms to deal with the same problem is because the training time for the first approach require plenty of time and memory, about 1-2 days for one cycle. Therefore, we need another method that shortens the training process. Both of these two approaches have their advantages and disadvantages. In the research, two different approaches for automatically generate tags are created. The two approaches are using different AI technique, and the results are generated and analyzed. The two approaches can work online. Our research demonstrate that using AI for generate tags for blog are feasible. The technique can be used in another area, e.g. generate HTML tags, document indexing. These areas of research have potential to continue because many researchers are keeping studying in this topic. Our research makes use of the collective intelligence to automatically generate tag suggestions to blog authors based on the semantic content of blog entries.|
|Online Catalog Link: ||http://lib.cityu.edu.hk/record=b2340803|
|Appears in Collections:||CS - Master of Philosophy |
Items in CityU IR are protected by copyright, with all rights reserved, unless otherwise indicated.