Abstract:
With the proliferation of social media sites, such
as Twitter, Facebook, and LinkedIn, social streams
have proven to contain the most up-to-date information
on current events. Therefore, it is crucial to extract
activities or events from the social streams, such as
tweets and it become an ongoing research trend. Most
approaches that aim at extracting event information
from twitter typically use the context of messages.
However, exploiting the location information of georeferenced
messages and the profile data are also
important because tweet messages are short,
fragmented and noisy, and therefore not include
complete information about the events. For this, in this
paper, a framework for event-extraction and
categorization from Twitter is proposed. To extract the
localized related activities, several mining mechanisms
and cleaning techniques is used for real-time twitter
corpus and various language processing approaches is
applied for categorization the events and then the
system will display the valuable information for the
targeted domain.
