Search

Rare diseases (RDs) are collectively common and affect 300 million people worldwide. Accurate phenotyping is critical for informing diagnosis and treatment, but RD phenotypes are often embedded in unstructured text and time-consuming to extract manually. While natural language processing (NLP) models can perform named entity recognition (NER) to automate extraction, a major bottleneck is the development of a large, annotated corpus for model training. Recently, prompt learning emerged as an NLP paradigm that can lead to more generalizable results without any (zero-shot) or few labeled samples (few-shot). Despite growing interest in ChatGPT, a revolutionary large language model capable of following complex human prompts and generating high-quality responses, none have studied its NER performance for RDs in the zero- and few-shot settings. To this end, we engineered novel prompts aimed at extracting RD phenotypes and, to the best of our knowledge, are the first the establish a benchmark for evaluating ChatGPT's performance in these settings. We compared its performance to the traditional fine-tuning approach and conducted an in-depth error analysis. Overall, fine-tuning BioClinicalBERT resulted in higher performance (F1 of 0.689) than ChatGPT (F1 of 0.472 and 0.591 in the zero- and few-shot settings, respectively). Despite this, ChatGPT achieved similar or higher accuracy for certain entities (i.e., rare diseases and signs) in the one-shot setting (F1 of 0.776 and 0.725). This suggests that with appropriate prompt engineering, ChatGPT has the potential to match or outperform fine-tuned language models for certain entity types with just one labeled sample. While the proliferation of large language models may provide opportunities for supporting RD diagnosis and treatment, researchers and clinicians should critically evaluate model outputs and be well-informed of their limitations.

Although geographically weighted Poisson regression (GWPR) is a popular regression for spatially indexed count data, its development is relatively limited compared to that found for linear geographically weighted regression (GWR), where many extensions (e.g., multiscale GWR, scalable GWR) have been proposed. The weak development of GWPR can be attributed to the computational cost and identification problem in the underpinning Poisson regression model. This study proposes linearized GWPR (L-GWPR) by introducing a log-linear approximation into the GWPR model to overcome these bottlenecks. Because the L-GWPR model is identical to the Gaussian GWR model, it is free from the identification problem, easily implemented, computationally efficient, and offers similar potential for extension. Specifically, L-GWPR does not require a double-loop algorithm, which makes GWPR slow for large samples. Furthermore, we extended L-GWPR by introducing ridge regularization to enhance its stability (regularized L-GWPR). The results of the Monte Carlo experiments confirmed that regularized L-GWPR estimates local coefficients accurately and computationally efficiently. Finally, we compared GWPR and regularized L-GWPR through a crime analysis in Tokyo.