دورية أكاديمية
أعرض في EDS

Clinical time series prediction: Toward a hierarchical dynamical system framework.

التفاصيل البيبلوغرافية
العنوان: Clinical time series prediction: Toward a hierarchical dynamical system framework.
المؤلفون: Liu Z; Computer Science Department, University of Pittsburgh, 210 South Bouquet Street, Pittsburgh, PA 15260, USA. Electronic address: ztliu@cs.pitt.edu., Hauskrecht M; Computer Science Department, University of Pittsburgh, 210 South Bouquet Street, Pittsburgh, PA 15260, USA.
المصدر: Artificial intelligence in medicine [Artif Intell Med] 2015 Sep; Vol. 65 (1), pp. 5-18. Date of Electronic Publication: 2014 Nov 06.
نوع المنشور: Journal Article; Research Support, N.I.H., Extramural
اللغة: English
بيانات الدورية: Publisher: Elsevier Science Publishing Country of Publication: Netherlands NLM ID: 8915031 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1873-2860 (Electronic) Linking ISSN: 09333657 NLM ISO Abbreviation: Artif Intell Med Subsets: MEDLINE
أسماء مطبوعة: Publication: Amsterdam : Elsevier Science Publishing
Original Publication: Tecklenburg, Federal Republic of Germany : Burgverlag, c1989-
مواضيع طبية MeSH: Data Mining* , Machine Learning* , Models, Statistical*, Clinical Decision-Making/*methods, Blood Cell Count ; Cardiac Surgical Procedures ; Computer Simulation ; Humans ; Normal Distribution ; Predictive Value of Tests ; Time Factors
مستخلص: Objective: Developing machine learning and data mining algorithms for building temporal models of clinical time series is important for understanding of the patient condition, the dynamics of a disease, effect of various patient management interventions and clinical decision making. In this work, we propose and develop a novel hierarchical framework for modeling clinical time series data of varied length and with irregularly sampled observations.
Materials and Methods: Our hierarchical dynamical system framework for modeling clinical time series combines advantages of the two temporal modeling approaches: the linear dynamical system and the Gaussian process. We model the irregularly sampled clinical time series by using multiple Gaussian process sequences in the lower level of our hierarchical framework and capture the transitions between Gaussian processes by utilizing the linear dynamical system. The experiments are conducted on the complete blood count (CBC) panel data of 1000 post-surgical cardiac patients during their hospitalization. Our framework is evaluated and compared to multiple baseline approaches in terms of the mean absolute prediction error and the absolute percentage error.
Results: We tested our framework by first learning the time series model from data for the patients in the training set, and then using it to predict future time series values for the patients in the test set. We show that our model outperforms multiple existing models in terms of its predictive accuracy. Our method achieved a 3.13% average prediction accuracy improvement on ten CBC lab time series when it was compared against the best performing baseline. A 5.25% average accuracy improvement was observed when only short-term predictions were considered.
Conclusion: A new hierarchical dynamical system framework that lets us model irregularly sampled time series data is a promising new direction for modeling clinical time series and for improving their predictive performance.
(Copyright © 2014 Elsevier B.V. All rights reserved.)
References: Artif Intell Med. 2000 Mar;18(3):221-44. (PMID: 10675716)
AMIA Annu Symp Proc. 2010 Nov 13;2010:286-90. (PMID: 21346986)
Artif Intell Med. 1995 Dec;7(6):515-40. (PMID: 8963374)
Nonlinear Dynamics Psychol Life Sci. 2006 Apr;10(2):187-214. (PMID: 16519865)
J Biomed Inform. 2013 Feb;46(1):47-55. (PMID: 22944172)
J Am Med Inform Assoc. 2014 May-Jun;21(3):501-8. (PMID: 24259520)
Proc AMIA Symp. 1998;:538-42. (PMID: 9929277)
Artif Intell Med. 2015 Sep;65(1):5-18. (PMID: 25534671)
Evolution. 1985 Jul;39(4):783-791. (PMID: 28561359)
Stud Health Technol Inform. 2010;160(Pt 2):861-5. (PMID: 20841808)
معلومات مُعتمدة: R01 GM088224 United States GM NIGMS NIH HHS; R01 LM010019 United States LM NLM NIH HHS; R01GM088224 United States GM NIGMS NIH HHS; R01LM010019 United States LM NLM NIH HHS
فهرسة مساهمة: Keywords: Clinical time series prediction; Gaussian processes; Hierarchical framework; Linear dynamical system
تواريخ الأحداث: Date Created: 20141224 Date Completed: 20160622 Latest Revision: 20240323
رمز التحديث: 20240323
مُعرف محوري في PubMed: PMC4422790
DOI: 10.1016/j.artmed.2014.10.005
PMID: 25534671
قاعدة البيانات: MEDLINE
الوصف
تدمد:1873-2860
DOI:10.1016/j.artmed.2014.10.005