ICTACT Journal on Soft Computing

S. Saraswathi ¹

Affiliations
1 Department of Information Technology, Pondicherry Engineering College, Pondicherry, IN

Abstract

This paper describes the design of online textual presentation tool developed by using various approaches like information retrieval, information extraction, topic identification and sentence weightage algorithms. To design a presentation for the given user query, the following approaches are used: Information extraction based on Boolean and probabilistic approaches, Topic identification based on semantic based approach, Sentence weightage algorithm based on importance of the terms and topics involved in the documents. A combination of these approaches led to the design of PowerPoint presentation file.

Keywords

Recall, Precision, Textual presentation, Topic Identification, Sentence Weightage, Summarization.

Full Text

     

S. Saraswathi ¹, R. Arti ²

Affiliations
1 Department of Information Technology, Pondicherry Engineering College, Pondicherry, IN
2 Microsoft R&D India Private Limited, Hyderabad, IN

Abstract

This paper investigates the use of clustering and lexical chains to produce coherent summaries of multiple documents in text format to generate an indicative, less redundant summary. The summary is designed as per user's requirement of conciseness i.e., the documents are summarized according to the percentage input by the user. For achieving the above, various clustering techniques are used. Clustering is done at two levels, one at single document level and then at multi-document level. The clustered sentences are scored based on five different methods and lexically linked to produce the final summary in a text document.

Keywords

Hierarchical Clustering, Lexical Chaining, Precision, Recall.

Full Text