Analyzing Content, Event, and Time by Temporal Collocations in Weblogs
Date Issued
2007
Date
2007
Author(s)
Teng, Chun-Yuan
DOI
en-US
Abstract
With the popularity of weblogs, it is desirable to extract abundant personal experiences, public opinions, and real events from weblogs. Although many researchers have analyzed the content of weblogs and real events, we do not find any works using multiword to discuss the relationship between the content and time. To enable the information retrieval of the content, time, and event, we provide several innovative techniques and algorithms to address these needs. (1) The temporal collocation is employed to observe the strength of term-to-term associations over time. (2) The event detection algorithm is to identify the collocations that may cause event in a specific timestamp. (3) The event description algorithm retrieves set of collocations which describe an event. In addition to these innovative techniques and algorithms, we also discuss the behavior of the temporal collocations and show the potential applications. The experimental results demonstrate that the temporal collocations capture the real world semantics and real world events over time. In general, the temporal collocations and the related techniques help users identify the real events and retrieve the interesting life patterns from weblogs.
Subjects
部落格
時間
內容
事件
自然語言
blog
weblog
collocation
time
temporal
content
event
detection
Type
thesis