Comprehensive Survey of Event Detection Techniques in Social Media Streams

Abstract

In the current years, the reception of net based more life stages are increased, for instance, Twitter, Facebook and their utilization as a component of the regular day to day existence of billions of people round the world. Given the habit of individuals to use these platforms to share thoughts, daily activities and experiences it's not stunning that the quantity of user generated content has reached unprecedented levels, with a considerable a part of that content being associated with real-world events.Event detection on social networking sites is crucial and important task in order to handle the emerging situations and close monitoring of the events. One of the major concern is to tackle the crisis like situations arising through this events. One of the way to control such activities is to detect the hot events on social networking sites as and when arises. Various detection models and techniques are available to detect events. This research focuses on surveying the various such detection models and techniques like Graph based algorithm, Multilayered inverted lists, CBOW model and Skip gram model, Incremental temporal topic model, Hypertext-Induced Topic Search (HITS) based Topic-Decision method (TD-HITS) algorithms, Multi-assignment graph partitioning algorithm. and comparing the performance of these models and algorithms in terms of their average accuracy. It is found that Hypertext-Induced Topic Search (HITS) based Topic-Decision method (TD-HITS) method gives the highest average accuracy and hence performs the best.This paper cover a detailed study of methods for event detection from Social Media that occurs over area and time. This paper highlight numerous event detection techniques and their limitations.