Spotted Hyena Optimization with Deep Learning-Based Automatic Text Document Summarization Model

Abstract

Automatic text summarization is an active investigation region determined as removing snippets or introductory sentences of a massive document and relating them as a short form of documents. Text Summarization can be either cost-efficient or time-efficient. An abstractive or extractive summary was studied with distinct algorithms comprising deep learning (DL), graph, and statistical-based techniques. DL has attained promising shows in comparison to the typical methods. With the development of various neural structures like the attention mechanism (usually called a transformer), there is a potential growth area for summarization tasks. Hence, this research presents a Spotted Hyena Optimization with Deep Learning based Automatic Text Summarization (SHODL-ATS) model. The SHODL-ATS technique's principal objective lies in the documents' automated summarization. To accomplish this, the presented SHODL-ATS technique performs data preprocessing to convert the data into a convenient form. The SHODL-ATS technique uses an Attention-based Bidirectional Gated Recurrent Unit (ABiGRU) model for summarizing the text documents. Finally, the SHO technique is enforced for the parameter tuning of the ABiGRU approach. To examine the achievement of the SHODL-ATS model, we validate the outcomes on benchmark datasets. The results indicate the promising achievement of the SHODL-ATS method over other existing techniques.