Keyword Based Tweet Extraction and Detection of Related Topics

Twitter is a micro blogging site that helps the transfer of information as short length tweets. The large quantum of information makes it necessary to find out methods and tools to summarize them. Our research work is to propose a method, which collect tweets using a specific keyword and then, summarizes them to find out topics related to that keyword. The topic detection is done by using clusters of frequent patterns. Already existing pattern oriented topic detection techniques suffer from the wrong correlation problem of patterns. In this paper, we propose two algorithms, TDA (Topic Detection using AGF) and TCTR (Topic Clustering and Tweet Retrieval), which will help to overcome this problem. From various experimental results, it is observed that the proposed method can maintain good performance irrespective of the size of the data set.