Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{190895,
        author = {Rajni and Rajneesh and Poonam Rani},
        title = {Clinical text summarization for smart health support system using self supervised machine learning model},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {9033-9041},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=190895},
        abstract = {Clinical documents such as discharge summaries, progress notes, and diagnostic reports contain essential patient information but are often lengthy, unstructured, and difficult for clinicians to interpret quickly. This creates challenges in timely decision-making within smart health support systems. To address this issue, this study proposes a self-supervised clinical text summarization model designed to automatically condense complex medical narratives while preserving factual accuracy and contextual meaning. The methodology involves using the MIMIC-III dataset comprising 265,000 clinical documents, which were preprocessed through de-identification, text cleaning, normalization, sentence segmentation, and tokenization. Summarization pairs were constructed using gold or heuristic-based silver targets. A self-supervised transformer was then trained with optimizations such as learning-rate warmup, dropout, gradient accumulation, mixed-precision training, and early stopping. Additional supervised models—including SVM, Random Forest, and a hybrid extractive–abstractive model—were evaluated for comparison. Results show that the proposed hybrid model achieved the highest accuracy of 94.8%, precision of 95%, recall of 93%, and F1-score of 94%, outperforming traditional models and demonstrating strong capability for generating clinically coherent and reliable summaries.},
        keywords = {Clinical Text Summarization; Self-Supervised Learning; Smart Health Support System; MIMIC-III Dataset; Transformer Model},
        month = {April},
        }