Template Based Abstractive Summarization of Twitter Topic with Speech Act

Now a days, people are using microblogging services such as Twitter, Facebook, Google+ etc. The content of such site is large number of small textual messages that is posted by millions of users, at random or in response to perceived events or situations. The top trending topics on Twitter.com each can have thousands of tweets. It is very time consuming and difficult attempts to read all the tweets under a particular Trending Topic.Automatic summarization of Twitter messages (tweets) is an urgent need for efficient processing of the tweeted information. Twitter topic summarization deals with short, dissimilar, and noisy nature of tweets. In this paper, we have used a speech act-guided summarization approach.For classification purpose we have used Bagging Ensemble approch with Naive Bayes Classifier. First we have to recognize the speech acts in tweets ,then we extract key words and phrases from the tweets. An extracted key terms are ranked and inserted into special summary templates designed for speech acts. Here, in this proposed work we have implemented Ngram selection algorithm to select top ranking keyword and phrases and then inserted into summary template.