Microblog Topic Detection Based on LDA Model and Single-Pass Clustering

Microblogging is a recent social phenomenon of Web2.0 technology, having applications in many domains. It is another form of social media, recognized as Real-Time Web Publishing, which has won an impressive audience acceptance and surprisingly changed online expression and interaction for millions of users.It is observed that clustering by topic can be very helpful for the quick retrieval of desired information. We propose a novel topic detection technique that permits to retrieve in real-time the most emergent topics expressed by the community. Traditional text mining techniques have no special considerations for short and sparse microblog data. Keeping in view these special characteristics of data, we adopt Single-pass Clustering technique by using Latent Dirichlet Allocation (LDA) Model in place of traditional VSM model, to extract the hidden microblog topics information. Experiments on actual dataset results showed that the proposed method decreased the probabilities of miss and false alarm, as well as reduced the normalized detection cost.