Automated multi-document summarization using extractive-abstractive approaches

Abstract

This study presents a multi-document text summarizing system that employs a  hybrid  approach,  including  both  extractive  and  abstractive  methods. The goal  of  document  summarizing  is  to  create  a  coherent  and  comprehensive summary that captures the essential information contained in the document.The  difficulty  in  multi-document  text  summarization  lies  in  the  lengthy nature  of  the  input  material  and  the  potential  for  redundant  information.  This study utilises a combination of methods to address this issue. This study uses the TextRank algorithm as an extractor for each document to condense the  input  sequence.  This  extractor  is  designed  to  retrieve  crucial  sentences from each document, which are then aggregated and utilised as input for the abstractor.  This  study  uses bidirectional  and  auto-regressive  transformers (BART)as  an  abstractor.  This  abstractor  serves  to  condense  the  primary sentences in each document into a more cohesive summary. The evaluation of this text summarizing system was conducted using the ROUGE measure. The  research   yields  ROUGE  R1  and  R2  scores  of  41.95  and  14.81, respectively.