Large language models versus classical machine learning performance in COVID-19 mortality prediction using high-dimensional tabular data

This study compared the performance of classical feature-based machine learning models (CMLs) and large language models (LLMs) in predicting COVID-19 mortality using high-dimensional tabular data from 9,134 patients across four hospitals. Seven CML models, including XGBoost and random forest (RF), were evaluated alongside eight LLMs, such as GPT-4 and Mistral-7b, which performed zero-shot classification on text-converted structured data. Additionally, Mistral-7b was fine-tuned using the QLoRA approach. XGBoost and RF demonstrated superior performance among CMLs, achieving F1 scores of 0.87 and 0.83 for internal and external validation, respectively. GPT-4 led the LLM category with an F1 score of 0.43, while fine-tuning Mistral-7b significantly improved its recall from 1% to 79%, yielding a stable F1 score of 0.74 during external validation. Although LLMs showed moderate performance in zero-shot classification, fine-tuning substantially enhanced their effectiveness, potentially bridging the gap with CML models. However, CMLs still outperformed LLMs in handling high-dimensional tabular data tasks. This study highlights the potential of both CMLs and fine-tuned LLMs in medical predictive modeling, while emphasizing the current superiority of CMLs for structured data analysis.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2025-11-28
DOI
https://doi.org/10.1038/s41598-025-26705-7
Citations
1
Primary Topic
COVID-19 diagnosis using AI
Type
article
Field-Weighted Citation Impact
1.36
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Large language models versus classical machine learning performance in COVID-19 mortality prediction using high-dimensional tabular data

Amirali Soheili, Hamidreza Hatamabadi, Saba Shahsavan, Hossein Majlesi et al.
1 citations
Scientific Reports
COVID-19 diagnosis using AI
1.36
article

Large language models versus classical machine learning performance in COVID-19 mortality prediction using high-dimensional tabular data

Amirali Soheili, Hamidreza Hatamabadi, Saba Shahsavan, Hossein Majlesi, Zohreh Tajabadi, Fatemeh Shojaeian, Zahra Atf, Hossein Toreyhi, Omid Yazdani, Ilad Alavi Darazam, Danial Samiei Nasr, Hossein Afshar, Girish N. Nadkarni, Jamil S. Samaan, Fatemeh Heydari, Mohamad Amin Pourhoseingholi, Erfan Taherifard, Naghmeh Asadimanesh, Mahshad Sarikhani, Amirreza Allahgholipour, Ankit Sakhuja, Thomas Savage, Ali Soroush, Mohammad Khalafi, Seyed Amir Ahmad Safavi‐Naini, Ali Moradi, Romina Esbati, Aryan Salahi‐Niri, Mohammadreza Ghaffarzadeh-Esfahani, Narges Gazmeh, Mahdi Ghaffarzadeh-Esfahani, Hadis Shabanipour, Alireza Soheilipour, Amirali Mohsenzadeh-Kermani, Moozhan Maleki, Ameneh Salehi, Sajjad Hossein Zadeh, Aydin Feyzi, Mohammad Hassan Bagheri, Farid Alimardani, Nooshin Goudarzi, Mohamadamin Tarighat-Payma
article en
1 citations

Abstract

This study compared the performance of classical feature-based machine learning models (CMLs) and large language models (LLMs) in predicting COVID-19 mortality using high-dimensional tabular data from 9,134 patients across four hospitals. Seven CML models, including XGBoost and random forest (RF), were evaluated alongside eight LLMs, such as GPT-4 and Mistral-7b, which performed zero-shot classification on text-converted structured data. Additionally, Mistral-7b was fine-tuned using the QLoRA approach. XGBoost and RF demonstrated superior performance among CMLs, achieving F1 scores of 0.87 and 0.83 for internal and external validation, respectively. GPT-4 led the LLM category with an F1 score of 0.43, while fine-tuning Mistral-7b significantly improved its recall from 1% to 79%, yielding a stable F1 score of 0.74 during external validation. Although LLMs showed moderate performance in zero-shot classification, fine-tuning substantially enhanced their effectiveness, potentially bridging the gap with CML models. However, CMLs still outperformed LLMs in handling high-dimensional tabular data tasks. This study highlights the potential of both CMLs and fine-tuned LLMs in medical predictive modeling, while emphasizing the current superiority of CMLs for structured data analysis.

Scientific ReportsVol. 15(1)
Isfahan University of Medical Sciences (IR), Cedars-Sinai Medical Center (US), Johns Hopkins University (US), Shiraz University of Medical Sciences (IR), Ontario Tech University (CA), Iranshahr University (IR), Research Institute for Endocrine Sciences (IR), Child Health and Development Institute (US), NIHR Nottingham Hearing Biomedical Research Unit (GB), Nottingham Biomedical Research Centre (GB), Imam Hossein Hospital (IR), Shahid Beheshti University of Medical Sciences (IR), Tehran University of Medical Sciences (IR), Stanford University (US), Icahn School of Medicine at Mount Sinai (US)
Openalex Percentile: Top 14%
COVID-19 diagnosis using AI
1.36
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.