Developing an Alert System Through Probabilistic Latent Semantic Indexing By Investigating Tweets

Nowadays, Social media such as Facebook, Twitter, you-tube, etc plays an important role to update instantly. Twitter is an Information sharing site, in which it is mainly used to share text message within 140 characters, audio links, video links, URL’s,.. It got more popular because of its real time nature. Users of this site frequently checks about what others doing and updates about “what’s happening?”. Because of its real time character , many applications are being developed. This paper is developed to alert the users by collecting the tweets and it must be compared with the training data set. Then, the obtained data is classified by a technique called SVM and Particle filtering algorithm is used to determine the “hot spot”(ie. the location in which event occurred). This paper presents an alert system to all the registered users for real time sensitive events like Earthquake, floods, storm ,etc by investigating small text messages(tweets).The main objective is to provide the alerts by few tweets through Probabilistic latent semantic indexing(PLSI) technique and reducing the false positive rate by re tweeting analysis.