Detecting Emergency Events and Geo-Location Awareness from Twitter Streams

Abstract

Abstract — the rapidly increasing number of messages on twitter is quite interesting. Through twitter streaming, this paper is capable of delivering tweets for any keywords from clients all around the world or Hashtag in real-time. However, semantic topic extraction and tracking the user-interested news events from messages on twitter can be considered as a challenging task. In this paper focused on detecting unusual behaviors and geo-social events in Twitter streams. The proposed method is able to provide/manage notifications and awareness (i.e., alerts) for users. In general, there are 4 steps in the process of identifying alerts and current awareness as follows: 1) Applying the set of real-time streaming APIs offered by Twitter for retrieving the public live stream of Twitter messages, 2) Utilizing an online clustering method to form the group of similar messages based on a corpus of text messages, 3) Applying an algorithm to compare the newly extracted event-term candidate with the list of event-term based on timestamp and calculate ranking indicating real-world events by frequent re-tweets, and 4) Identifying the alerts and current awareness based on the user's geo-location. Into the bargain, the proposed method is evaluated using high-volume Twitter data and as a result the findings are reported.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 8 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.