Methods for selection of adequate neural network structures with application to early assessment of chest pain patients by biochemical monitoring

被引:10
作者
Ellenius, J [1 ]
Groth, T [1 ]
机构
[1] Uppsala Univ, Dept Med Sci, Unit Biomed Informat & Syst Anal, S-75185 Uppsala, Sweden
关键词
artificial neural network; ANN identifiability; acute myocardial infarction; myoglobin; creatine kinase MB; troponin T;
D O I
10.1016/S1386-5056(00)00065-4
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
A methodology for selecting, training and estimating the performance of adequate artificial neural network (ANN) structures and incorporating them with algorithms that are optimized for clinical decision making is presented. The methodology was applied to the problem of early ruling-in/ruling-out of patients with suspected acute myocardial infarction using frequent biochemical monitoring. The selection of adequate ANN structures from a set of candidates was based on criteria for model compatibility, parameter identifiability and diagnostic performance. The candidate ANN structures evaluated were the single-layer perceptron (SLP), the fuzzified SLP, the multiple SLP, the gated multiple SLP, the multi-layer perceptron (MLP) and the discrete-time recursive neural network. The identifiability of the ANNs was assessed in terms of the conditioning of the Hessian of the objective function, and variability of parameter estimates and decision boundaries in the trials of leave-one-out cross-validation. The commonly used MLP was shown to be non-identifiable for the present problem and available amount of data, despite artificially reducing the model complexity with use of regularization methods. The investigation is concluded by recommending a number of guidelines in order to obtain an adequate ANN model. (C) 2000 Elsevier Science Ireland Ltd. All rights reserved.
引用
收藏
页码:181 / 202
页数:22
相关论文
共 33 条
[1]  
ALBERTINI F, 1992, P 31 IEEE C DEC CONT, P3822
[2]  
Anderson H, 1991, Lakartidningen, V88, P2890
[3]  
ANDERSON JA, 1982, LOGISTIC DISCRIMINAT, P169
[4]  
*BIOMACS STUD GROU, 1995, CORON ARTERY DIS, V6, P321
[5]  
Bishop C. M., 1995, NEURAL NETWORKS PATT
[6]  
Clopper CJ, 1934, BIOMETRIKA, V26, P404, DOI 10.2307/2331986
[7]   COMPUTATIONAL INTELLIGENCE FOR LABORATORY INFORMATION-SYSTEMS [J].
EKLUND, P ;
FORSSTROM, JJ .
SCANDINAVIAN JOURNAL OF CLINICAL & LABORATORY INVESTIGATION, 1995, 55 :21-30
[8]  
Ellenius J, 1997, CLIN CHEM, V43, P1919
[9]   USING DATA PREPROCESSING AND SINGLE-LAYER PERCEPTRON TO ANALYZE LABORATORY DATA [J].
FORSSTROM, JJ ;
IRJALA, K ;
SELEN, G ;
NYSTROM, M ;
EKLUND, P .
SCANDINAVIAN JOURNAL OF CLINICAL & LABORATORY INVESTIGATION, 1995, 55 :75-81
[10]   NEURAL NETWORK ANALYSIS OF SERIAL CARDIAC ENZYME DATA - A CLINICAL-APPLICATION OF ARTIFICIAL MACHINE INTELLIGENCE [J].
FURLONG, JW ;
DUPUY, ME ;
HEINSIMER, JA .
AMERICAN JOURNAL OF CLINICAL PATHOLOGY, 1991, 96 (01) :134-141