Keyword Identification and Summarization in a Text by Using Deep Learning Based Models

The growing volume of textual data across various domains has led to the increasing demand for efficient and accurate text processing systems. This research proposes a transformer-based pipeline that integrates DistilBERT for contextual keyword extraction and T5 for abstractive text summarization. The proposed system effectively identifies meaningful keywords while generating concise, coherent, and contextually accurate summaries. The model is evaluated using benchmark datasets and standard metrics such as Precision, Recall, F1-score, and ROUGE, demonstrating superior performance over traditional extractive summarization techniques. The experimental results indicate that the system enhances the relevance and readability of summaries, making it suitable for applications in journalism, healthcare, legal document analysis, and academic research. Furthermore, the pipeline efficiently handles large text, reducing the manual effort required for text comprehension and information retrieval.