Text Summarization using NLP Technique
This research analysis uses the BBC news dataset to evaluate and compare the results obtained from the machine learning models and transformer architecture-based pre-trained models.
Abstract
The text data online is increasing massively; hence, producing a summarized text document is essential. We can create the summarization of multiple text documents either manually or automatically. A manual approach may be tedious and a time-consuming process. The resulting composition may not be accurate when processing lengthy articles; hence the second approach, i.e., the automated summary generation process, is essential. Training machine learning models using these processes makes space and time-efficient summary generation possible. There are two widely used methods to generate summaries, namely, Extractive summarization and abstractive summarization. The extractive technique scans the original document to find the relevant sentences and extracts only that information from it. The abstractive summarization technique interprets the original text before generating the summary. This process is more complicated, and transformer architecture-based pre-trained models are used for comparing the text & developing the outline. This research analysis uses the BBC news dataset to evaluate and compare the results obtained from the machine learning models.