Using Connectionist Modules for Decision Support

Abstract

A connectionist model for decision support was constructed out of several back-propagation modules. Manifestations serve as input to the model; they may be real-valued, and the confidence in their measurement may be specified. The model produces as its output the posterior probability of disease. The model was trained on 1,000 cases taken from a simulated underlying population with three conditionally independent manifestations. The first manifestation had a linear relationship between value and posterior probability of disease, the second had a stepped relationship, and the third was normally distributed. An independent test set of 30,000 cases showed that the model was better able to estimate the posterior probability of disease (the standard deviation of residuals was 0.046, with a 95% confidence interval of 0.046-0.047) than a model constructed using logistic regression (with a standard deviation of residuals of 0.062, with a 95% confidence interval of 0.062-0.063). The model fitted the normal and stepped manifestations better than the linear one. It accommodated intermediate levels of confidence well.

Key-Words

Connectionist Models - Neural Networks - Confidence - Decision Support - Logistic Regression

• REFERENCES

• 1 Mulsant BH. A neural network as an approach to clinical diagnosis. MD Computing 1990; 07: 25-36.
• 2 Hudson DL, Cohen ME, Anderson ME. Determination of testing efficacy in carcinoma of the lung using a neural network model. In: Greenes RA. ed. Proc Twelfth Annual Symposium on Computer Applications in Medical Care. New York: IEEE Computer Society Press; 1988: 251-5.
• 3 Fogelman FSoulie, Gallinari P, Le Cun Y, Thiria S. Network learning. To appear. In: Machine Learning, Vol.3.
• 4 Smith JW, Everhart JE, Dickson WC, Knowler WC, Johannes RS. Using the ADAP learning algorithm to forecast the onset of diabetes mellitus. In: Greenes RA. ed. Proc Twelfth Annual Symposium on Computer Applications in Medical Care. New York: IEEE Computer Society Press; 1988: 261-5.
• 5 Hripcsak G. The estimation of posterior probabilities using neural networks. In: Greenes RA. ed. Proc Twelfth Annual Symposium on Computer Applications in Medical Care. New York: IEEE Computer Society Press; 1988: 240-4.
• 6 Hripcsak G. A Neural Network for Medical Diagnosis that Accommodates Unreliable Data. In: Hammond WE. ed. Proc American Association for Medical Systems and Informatics Congress 89. Washington DC: AAMSI Publishers; 1989: 159-63.
• 7 Cohen ME, Hudson DL, Anderson MF. Development of a decision making model for carcinoma of the lung. In: Hammond WE. ed. Proc American Association for Medical Systems and Informatics Congress 89. Washington DC: AAMSI Publishers; 1989: 164-9.
• 8 Gallant SI. Connectionist expert systems. ’ Communications of the ACM 1988; 31: 152-69.
• 9 Wald J, Farach M, Tagamets M, Reggia JA. Generating plausible diagnostic hypotheses with self-processing causal networks. To appear. In: J of Experimental and Theoretical AI.
• 10 D’Autrechy CL, Reggia JA, Sutton III GG, Goodall SM. Maryland MIRRORS/II: a connectionist model simulator. MD Computing 1990; 07: 12-24.
• 11 Yoon YO, Peterson LL. DESKNET: The dermatology expert system with knowledge-based network. In: Amari S-I, Gross-berg S, Kohonen T. eds. Abstracts First Annual International Neural Network Society Meeting. Elmsford NY: Pergamon Press Inc; 1988: 477.
• 12 Ramamoorthy PA, Ho S. A neural network approach for implementing expert systems. In: Amari S-L, Grossberg S, Kohonen T. eds. Abstracts First Annual International Neural Network Society Meeting. Elmsford NY: Pergamon Press Inc; 1988: 44.
• 13 Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic Research: Principles and Quantitative Methods. Belmont CA: Lifetime Learning Publications; 1982
• 14 Afifi AA, Clark V. Computer-Aided Multivariate Analysis. Belmont CA: Lifetime Learning Publications; 1984
• 15 Breiman L, Friedman JH, Olshen RA, Stone CJ. Classification and Regression Trees. Monterey CA: Wadsworth and Brooks; 1984
• 16 Adlassnig K-P. Computer-based diagnostic screening and consultation in hospital. In: Salamon R, Blum B, Jorgensen M. eds. Medinfo 86. Amsterdam: North-Holland Pub Comp, y1986. 101-5.
• 17 Zadeh LA. The concept of a linguistic variable and its application to approximate reasoning – I. Information Sciences 1975; 08: 199-249.
• 18 Adlassnig K-P. A fuzzy logical model of computer-assisted medical diagnosis. Meth Inform Med 1980; 19: 141-8.
• 19 Hudson DL, Cohen ME. The impact of reasoning with uncertainty in a medical expert system. In: Stead WW. ed. Proc Eleventh Annual Symposium on Computer Applications in Medical Care. New York: IEEE Computer Society Press; 1987: 43-8.
• 20 Warner HR. Computer-Assisted Medical Decision-Making. New York NY: Academic Press; 1979: 13.
• 21 Szolovits P, Pauker SG. Categorical and probabilistic reasoning in medical diagnosis. Artificial Intelligence 1978; 11: 115-44.
• 22 Pearl J. Fusion, propagation, and structuring in belief networks. Artificial Intelligence 1986; 29: 241-88.
• 23 Buchanan BG, Shortliffe EH. Rule-Based Expert Systems: The MYCIN Experiments of the Stanford Heuristic Programming Project. Reading MA: Addison-Wesley Publishing Company; 1984
• 24 Kyburg HE. Bayesian and non-bayesian evidential updating. Artificial Intelligence 1987; 31: 271-93.
• 25 Kors JA, Talmon JL, van Bemmel JH. Multilead ECG Analysis. Comp Biomed Res 1986; 19: 28-46.
• 26 Rumelhart DE, McClelland JL. Parallel Distributed Processing (Vols. 1 and 2).. Cambridge MA: MIT Press; 1986
• 27 Hripcsak G. Toward data analysis using connectionist modules. In: Kingsland LC. ed. Proc Thirtheenth Annual Symposium on Computer Applications in Medical Care. New York: IEEE Computer Society Press; 1989: 271-5.
• 28 Hinton GE. Connectionist learning procedures. To appear. In: Artificial Intelligence.
• 29 Davies TR. Some notes on the probabilistic semantics of logistic function parameters in neural networks. In: Amari S-I, Grossberg S, Kohonen T. eds. Abstracts First Annual International Neural Network Society Meeting. Elmsford NY: Pergamon Press Inc; 1988: 88.
• 30 McClelland JL, Rumelhart DE. Explorations in Parallel Distributed Processing: A Handbook of Modules, Programs, and Exercises. Cambridge MA: The MIT Press; 1988
• 31 Snedecor GW, Cochran WG. Statistical Methods. 7th ed. Ames IA: The Iowa State University Press; 1980