Monitoring System for Traffic Analysis Using Twitter Stream

Abstract

Social networks are often utilized as a supply of data for event detection like road holdup and automobile accidents. Existing system present a period of time observance system for traffic event detection from twitter. The system fetches tweets from twitter and then; processes tweets victimisation text mining techniques. Last performs the classification of tweets. The aim of the system is to assign the suitable category label to every tweet, whether or not it's associated with a traffic event or not. System utilized the support vector machine as a classification model. The projected system uses the system supported semi-supervised approach, which provides coaching victimisation traffic connected dataset. we have a tendency to propose a bunch approach for classification of the tweets in traffic connected and non- traffic connected tweets. We use a geometer distance to calculate the similarity between the tweets.