MTLNFM: A Multi-Task Framework Using Neural Factorization Machines to Predict Patient Clinical Outcomes

Accurately predicting patient clinical outcomes is a complex task that requires integrating diverse factors, including individual characteristics, treatment histories, and environmental influences. This challenge is further exacerbated by missing data and inconsistent data quality, which often hinde...

Full description

Saved in:
Bibliographic Details
Main Authors: Rui Yin, Jiaxin Li, Qiang Yang, Xiangyu Chen, Xiang Zhang, Mingquan Lin, Jiang Bian, Ashwin Subramaniam
Format: Article
Language:English
Published: MDPI AG 2025-08-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/15/8733
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849239772497707008
author Rui Yin
Jiaxin Li
Qiang Yang
Xiangyu Chen
Xiang Zhang
Mingquan Lin
Jiang Bian
Ashwin Subramaniam
author_facet Rui Yin
Jiaxin Li
Qiang Yang
Xiangyu Chen
Xiang Zhang
Mingquan Lin
Jiang Bian
Ashwin Subramaniam
author_sort Rui Yin
collection DOAJ
description Accurately predicting patient clinical outcomes is a complex task that requires integrating diverse factors, including individual characteristics, treatment histories, and environmental influences. This challenge is further exacerbated by missing data and inconsistent data quality, which often hinder the effectiveness of traditional single-task learning (STL) models. Multi-Task Learning (MTL) has emerged as a promising paradigm to address these limitations by jointly modeling related prediction tasks and leveraging shared information. In this study, we proposed MTLNFM, a multi-task learning framework built upon Neural Factorization Machines, to jointly predict patient clinical outcomes on a cohort of 2001 ICU patients. We designed a preprocessing strategy in the framework that transforms missing values into informative representations, mitigating the impact of sparsity and noise in clinical data. We leveraged the shared representation layers, composed of a factorization machine and dense neural layers that can capture high-order feature interactions and facilitate knowledge sharing across tasks for the prediction. We conducted extensive comparative experiments, demonstrating that MTLNFM outperforms STL baselines across all three tasks (i.e., frailty status, hospital length of stay and mortality prediction), achieving AUROC scores of 0.7514, 0.6722, and 0.7754, respectively. A detailed case analysis further revealed that MTLNFM effectively integrates both task-specific and shared representations, resulting in more robust and realistic predictions aligned with actual patient outcome distributions. Overall, our findings suggest that MTLNFM is a promising and practical solution for clinical outcome prediction, particularly in settings with limited or incomplete data, and can support more informed clinical decision-making and resource planning.
format Article
id doaj-art-e73fae0d4c9f496aaf6495503dd7ab8b
institution Kabale University
issn 2076-3417
language English
publishDate 2025-08-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-e73fae0d4c9f496aaf6495503dd7ab8b2025-08-20T04:00:50ZengMDPI AGApplied Sciences2076-34172025-08-011515873310.3390/app15158733MTLNFM: A Multi-Task Framework Using Neural Factorization Machines to Predict Patient Clinical OutcomesRui Yin0Jiaxin Li1Qiang Yang2Xiangyu Chen3Xiang Zhang4Mingquan Lin5Jiang Bian6Ashwin Subramaniam7Department of Health Outcomes and Biomedical Informatics, University of Florida, Gainesville, FL 32610, USASchool of Electrical and Electronic Engineering, Nanyang Technological University of Singapore, Singapore 639798, SingaporeDepartment of Health Outcomes and Biomedical Informatics, University of Florida, Gainesville, FL 32610, USASchool of Computer Science and Engineering, University of New South Wales, Kensington, NSW 2052, AustraliaCollege of Computing and Informatics, University of North Carolina Charlotte, Charlotte, NC 28223, USADivision of Computational Health Sciences, University of Minnesota, Minneapolis, MN 55455, USADepartment of Biostatistics and Health Data Science, Indiana University, Bloomington, IN 47405, USAMonash Health, Monash University, Melbourne, VIC 3800, AustraliaAccurately predicting patient clinical outcomes is a complex task that requires integrating diverse factors, including individual characteristics, treatment histories, and environmental influences. This challenge is further exacerbated by missing data and inconsistent data quality, which often hinder the effectiveness of traditional single-task learning (STL) models. Multi-Task Learning (MTL) has emerged as a promising paradigm to address these limitations by jointly modeling related prediction tasks and leveraging shared information. In this study, we proposed MTLNFM, a multi-task learning framework built upon Neural Factorization Machines, to jointly predict patient clinical outcomes on a cohort of 2001 ICU patients. We designed a preprocessing strategy in the framework that transforms missing values into informative representations, mitigating the impact of sparsity and noise in clinical data. We leveraged the shared representation layers, composed of a factorization machine and dense neural layers that can capture high-order feature interactions and facilitate knowledge sharing across tasks for the prediction. We conducted extensive comparative experiments, demonstrating that MTLNFM outperforms STL baselines across all three tasks (i.e., frailty status, hospital length of stay and mortality prediction), achieving AUROC scores of 0.7514, 0.6722, and 0.7754, respectively. A detailed case analysis further revealed that MTLNFM effectively integrates both task-specific and shared representations, resulting in more robust and realistic predictions aligned with actual patient outcome distributions. Overall, our findings suggest that MTLNFM is a promising and practical solution for clinical outcome prediction, particularly in settings with limited or incomplete data, and can support more informed clinical decision-making and resource planning.https://www.mdpi.com/2076-3417/15/15/8733multi-task learningneural factorization machinemachine learningclinical outcome prediction
spellingShingle Rui Yin
Jiaxin Li
Qiang Yang
Xiangyu Chen
Xiang Zhang
Mingquan Lin
Jiang Bian
Ashwin Subramaniam
MTLNFM: A Multi-Task Framework Using Neural Factorization Machines to Predict Patient Clinical Outcomes
Applied Sciences
multi-task learning
neural factorization machine
machine learning
clinical outcome prediction
title MTLNFM: A Multi-Task Framework Using Neural Factorization Machines to Predict Patient Clinical Outcomes
title_full MTLNFM: A Multi-Task Framework Using Neural Factorization Machines to Predict Patient Clinical Outcomes
title_fullStr MTLNFM: A Multi-Task Framework Using Neural Factorization Machines to Predict Patient Clinical Outcomes
title_full_unstemmed MTLNFM: A Multi-Task Framework Using Neural Factorization Machines to Predict Patient Clinical Outcomes
title_short MTLNFM: A Multi-Task Framework Using Neural Factorization Machines to Predict Patient Clinical Outcomes
title_sort mtlnfm a multi task framework using neural factorization machines to predict patient clinical outcomes
topic multi-task learning
neural factorization machine
machine learning
clinical outcome prediction
url https://www.mdpi.com/2076-3417/15/15/8733
work_keys_str_mv AT ruiyin mtlnfmamultitaskframeworkusingneuralfactorizationmachinestopredictpatientclinicaloutcomes
AT jiaxinli mtlnfmamultitaskframeworkusingneuralfactorizationmachinestopredictpatientclinicaloutcomes
AT qiangyang mtlnfmamultitaskframeworkusingneuralfactorizationmachinestopredictpatientclinicaloutcomes
AT xiangyuchen mtlnfmamultitaskframeworkusingneuralfactorizationmachinestopredictpatientclinicaloutcomes
AT xiangzhang mtlnfmamultitaskframeworkusingneuralfactorizationmachinestopredictpatientclinicaloutcomes
AT mingquanlin mtlnfmamultitaskframeworkusingneuralfactorizationmachinestopredictpatientclinicaloutcomes
AT jiangbian mtlnfmamultitaskframeworkusingneuralfactorizationmachinestopredictpatientclinicaloutcomes
AT ashwinsubramaniam mtlnfmamultitaskframeworkusingneuralfactorizationmachinestopredictpatientclinicaloutcomes