Improved Cuckoo Search Optimization Algorithm based Multi-document Summarization Model

摘要

Text summarization is a procedure of extracting the meaning details from source documents. Multi-document summarization is an automated task where the necessary details are extracted using several input documents. Diverse techniques were presented on the generation of the summary from single or many documents. Single and multiple document summarization techniques pose several issues because they might comprise of maximum repetitive information compared to the single document. To resolve this issue, this study develops a novel improved cuckoo search optimization algorithm (ICSA) for multi-document summarization. For improving the optimization abilities of the CS technique, the orthogonal concepts and simulated annealing (SA) are included to raise the exploitation searching capability. The presented ICSA model involves distinct sub-processes namely preprocessing, input, and summary representation. The main aim is to summarize the representation task involves the creation of a document summary including the meaning data. Through the optimal sentence choosing task using the ICSA model, the required sentences defining the summary are selected. An extensive experimental validation process was carried out on the DUC 2006 and 2007 datasets and the efficiency is examined in terms of distinct aspects.