English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Background: Encouraged by the success of pretrained Transformer models inmany natural language processing tasks, their use for InternationalClassification of Diseases (ICD) coding tasks is now actively being explored.In this study, we investigate three types of Transformer-based models, aimingto address the extreme label set and long text classification challenges thatare posed by automated ICD coding tasks. Methods: The Transformer-based modelPLM-ICD achieved the current state-of-the-art (SOTA) performance on the ICDcoding benchmark dataset MIMIC-III. It was chosen as our baseline model to befurther optimised. XR-Transformer, the new SOTA model in the general extrememulti-label text classification domain, and XR-LAT, a novel adaptation of theXR-Transformer model, were also trained on the MIMIC-III dataset. XR-LAT is arecursively trained model chain on a predefined hierarchical code tree withlabel-wise attention, knowledge transferring and dynamic negative samplingmechanisms. Results: Our optimised PLM-ICD model, which was trained with longertotal and chunk sequence lengths, significantly outperformed the current SOTAPLM-ICD model, and achieved the highest micro-F1 score of 60.8%. TheXR-Transformer model, although SOTA in the general domain, did not perform wellacross all metrics. The best XR-LAT based model obtained results that werecompetitive with the current SOTA PLM-ICD model, including improving themacro-AUC by 2.1%. Conclusion: Our optimised PLM-ICD model is the new SOTAmodel for automated ICD coding on the MIMIC-III dataset, while our novel XR-LATmodel performs competitively with the previous SOTA PLM-ICD model.

Automated ICD Coding using Extreme Multi-label Long Text Transformer-based Models