Text Summarization using Neural Networks
Anish Jadhav, Rajat Jain, Steve Fernandes, Sana Shaikh · 2019
Text Summarization is the technique of extricating notable data from the first content archive. In this procedure, the separated data is produced as a consolidated report and introduced as a clearly expressed rundown. Text Summarization can be broadly classified into Extractive Summarization and Abstractive Summarization. This paper describes an algorithm based on a combination of both these approaches. Initially, essential sentences are identified and stitched together to form a consolidated report. The significance of a sentence is chosen in light of measurable and semantic highlights of sentences. This shorter representation is then passed through an Encoder-Decoder model to generate a concise summary representing the whole article. The proposed model is capable of effectively creating a concise summary, which is semantically and linguistically correct, by understanding the whole content and reletting it in its own words. The proposed methodology focuses only on the relevant sentences and passes it to the Bi-Directional RNN for identifying and representing the core idea of the article.