Title: Error tolerant global search incorporated with deep learning algorithm to automatic Hindi text summarisation
Authors: J. Anitha; P.V.G.D. Prasad Reddy; M.S. Prasad Babu
Addresses: Vignan's Institute of Information Technology, Duvvada, Vishakhapatnam, India ' Andhra University, Vishakapatnam, India ' Andhra University, Vishakapatnam, India
Abstract: There is an exponential growth in the available electronic information in the last two decades. It causes a huge necessity to quickly understand high volume text data. This paper describes an efficient algorithm and it works by assigning scores to sentences in the document which is to be summarised. It also focuses on document extracts; a particular kind of computed document summary. The proposed approach uses fuzzy classifier and deep learning algorithm. Fuzzy classifier produces score for each sentence and the deep learning (DL) also produces score for each sentence. The combination of score from both fuzzy classifier and DL produces the hybrid score. Finally, the summarised text can be generated based on this hybrid score. In our proposed approach, we have achieved an average precision rate of 0.92 and average recall rate of 0.88 and the compression rate is 10% according to the experimental analysis.
Keywords: GSA; fuzzy; summarisation; hybrid; deep learning; DL.
DOI: 10.1504/IJBIDM.2019.098841
International Journal of Business Intelligence and Data Mining, 2019 Vol.14 No.3, pp.359 - 380
Received: 21 Jan 2017
Accepted: 06 Mar 2017
Published online: 04 Apr 2019 *