A developed framework for multi-document summarization using softmax regression and spider monkey optimization methods

In everyday life, multi-document summarization (MDS) methods are becoming tremendous attention in different fields, especially for online documents, because this online document conveys information to users by generating succinct and comprehensive summary. The summarized document contains the summary of various documents with same topic. Here, Spider Monkey Optimization (SMO) algorithm is introduced for summary generation. Before that, multi-documents are compressed into single document and different pre-processing methods are applied to remove the unwanted word from the document. Then, semantic and syntactic features are extracted from the document using different methods. The mined features are then provided into the softmax regression (SR) technique for further processing. Finally, SMO algorithm is proposed to generate the summary about whole document. The proposed text summarization process is implemented in Python platform using the BBC news dataset, DUC (Document Understanding Conference) 2002, 2006, and 2007 datasets. During pre-processing, the tokenization is performed by Natural Language Tool Kit (NLTK) tool and the lemmatization in WordNet lemmatizer. The terms recall, F-measure and precision are offered in this work for performance evaluation, and the accuracy of this method is found to be better than the other existing MDS techniques.