Odnn: An Automatic Optimized Deep Neural Network Based Multi-Document Summarization
Abstract
As the information on the web is growing rapidly, automatic multi-document summarization plays the vital role for understanding the information in short about the documents. However, it is difficult to analyze about a reasonable and specific document related with a particular subject within a limited time period. So, in this paper, we present an automatic optimized deep neural network (ODNN) based multi-document summarization. The proposed summarization approach follows the five phases that are i) pre-processing, ii) data conversion, iii) feature extraction, iv) sentence score calculation and v) rank based summarization. For sentence score calculation, we present ODNN in which weight parameters are selected optimally using pigeon optimization algorithm (POA). Depend on the score, the sentences are ranked and summarized. The performance of the proposed approach is analyzed using DUC 2002 dataset and is evaluated in terms of precision, recall and F-measure.