A novel extractive multi-document text summarization system using quantum-inspired genetic algorithm: MTSQIGA

The explosive growth of textual data on the web and the problem of obtaining desired information through this enormous volume of data has led to a dramatic increase in demand for developing automatic text summarization systems. For this reason, this paper presents a novel multi-document text summarization approach, called MTSQIGA, which extracts salient sentences from source document collection to generate the summary. The proposed generic summarizer models extractive summarization as a binary optimization problem that applies a modified quantum-inspired genetic algorithm (QIGA) in its processing stage to find the best solution. Objective function of our approach plays an important role in optimizing linear combination of coverage, relevance, and redundancy factors which consists of six sentence scoring measures. To ensures the generation of a summary with predefined length limit, the presented QIGA employs a modified quantum measurement and a self-adaptive quantum rotation gate based on the quality and length of the summary. Evaluation of the proposed system was performed on DUC 2005 and 2007 benchmark datasets in terms of ROUGE standard measures. Comparison of MTSQIGA with existing state-of-the-art approaches for multi-document summarization shows superior performance of the proposed systems over other methods on both existing benchmark datasets. It also indicates promising efficiency of our proposed algorithm on applying quantum-inspired genetic algorithm to the text summarization tasks.