Extractive text summarization for scientific journal articles using long short-term memory and gated recurrent units

Along with the increasing number of scientific publications, many communities must read entire text to get essence information from a journal article. This will be quite inconvenient if article is long and there are more than one journals. Motivated by this problem, encourages need for method summarization that can automatically, concisely, accurately summarize document. The purpose research create an extractive doing feature engineering extract semantic original text. Comparing short-term memory algorithm gated recurrent units were used most relevant sentences served as summary. results showed both algorithms yielded relatively similar accuracy results, at 98.40% 98.68%. evaluation matrix recall-oriented understudy gisting (ROUGE) evaluate summary results. produced LSTM model compared using latent analysis (LSA) then obtained recall values ROUGE-1, ROUGE-2, ROUGE-L respectively 76.25%, 59.49%, 72.72%.