Authors : Mrunal S. Bewoor and Suhas H. Patil
Abstract: Availability of huge amount of text data and increase of organizational spread over has arises the need to control their data corpora, especially with the availability of big data platforms. People does not have sufficient time to read and understand each document to make decisions based on document content. This has resulted in a great demand to summarize text documents to provide the end user a representative substitute for the original text input. This arises a need to identify techniques that performs precised summary retrieval through search queries against input documents. The user expects this process in a optimum way. To improve this process of querying against the full spectrum of original documents several generic algorithmms for text summarization have been developed, each with its own advantages and disadvantages. The study conducts a survey and analysis of the cluster based summary techniques obtained through expectation maximization, DBSCAN, graph based method, hierarchical and fuzzy C-means clustering algorithms. The results of the summaries obtained using these algorithms are evaluated with the parameters precision, recall, F-measure, compression ratio and retention ratio. The study aims at the analysis, investigation, design and development of various metrics which may help the end user regarding the selection of optimal query based technique.
Mrunal S. Bewoor and Suhas H. Patil, 2017. Analysis of Cluster Based Document Condensation Techniques. Journal of Engineering and Applied Sciences, 12: 5533-5536.