An Adaptive Mechanism For Event Detection in Twitter Using Big Data

The amount of digital information, such as text and photos, that is freely available online has increased nowadays, resulting in new insights and research opportunities through new channels. Text mining has sparked a lot of interest in the field of big data. In recent years, Twitter has grown in popularity as a social networking service. Twitter is being used by users to post on real-life activities, which reflect in many events categories. There is a need of developing the most scalable mechanism that can categorize a large number of events when supplied with huge social media tweets. We aim to build an efficient detection model that can segregate different types of events into their respective categories and they can be further used for Analysis and estimating the next series of events that can happen as a consequence. This project concentrates on the previous work done in this area and a new approach that can be built using Transfer Learning & Deep Learning with Apache Spark on text and image. This work speaks about the adaptive approach for the events that can be detected in the given big data. By comparing various existing detection mechanisms we show greater accuracy in our methodology which helps in effective identification. Explore different mechanisms for detection and how the considered adaptive approach can create a difference in achieving effective detection of text and image-related events when supplied at a huge rate. Here we are going to consider the environment, Transport, Geo-spatial, Water, and Education categories of events and would like to show the applicability of deep Learning at the classification phase.