3
© Modeling Obesity Using Abductive Networks Abdel-Aal, RE; Mangoud, AM ACADEMIC PRESS INC, COMPUTERS AND BIOMEDICAL RESEARCH; pp: 451-471; Vol: 30 King Fahd University of Petroleum & Minerals http://www.kfupm.edu.sa Summary er investigates the use of abductive-network machine learning for modeling icting outcome parameters in terms of input parameters in medical survey re we consider modeling obesity as represented by the waist-to-hip ratio isk factor to investigate the influence of various parameters. The would be useful in predicting values of clinical parameters that are difficu nsive to measure from others that are more readily available. The e network machine learning tool was used to model the WHR from 13 other arameters. Survey data were collected for a randomly selected sample of 1100 aged 20 yr and over attending nine primary health care centers at Al-Khobar, abia. Models were synthesized by training on a randomly selected set of 800 sing both continuous and categorical representations of the parameters, and d by predicting the WHR value for the remaining 300 cases. Models for continuous variable predict the actual values within an error of 7.5% at the idence limits. Categorical models predict the correct logical value of WHR error in only 2 of the 300 evaluation cases. Analytical relationships derived mple categorical models explain global observations on the total sur on to an accuracy as high as 99%. Simple continuous models represented as al functions highlight global relationships and trends. Results confirm orrelation between WHR and diastolic blood pressure, cholesterol level, and history of obesity. Compared to other statistical and neural es, AIM abductive networks provide faster and more automated model s. A review is given of other areas where the proposed modeling approach can l in clinical practice. (C) 1997 Academic Press. Copyright: King Fahd University of Petroleum & Minerals; http://www.kfupm.edu.sa

© Modeling Obesity Using Abductive Networks Abdel-Aal, RE; Mangoud, AM ACADEMIC PRESS INC, COMPUTERS AND BIOMEDICAL RESEARCH; pp: 451-471; Vol: 30 King

Embed Size (px)

Citation preview

Page 1: © Modeling Obesity Using Abductive Networks Abdel-Aal, RE; Mangoud, AM ACADEMIC PRESS INC, COMPUTERS AND BIOMEDICAL RESEARCH; pp: 451-471; Vol: 30 King

©

Modeling Obesity Using Abductive Networks

Abdel-Aal, RE; Mangoud, AM

ACADEMIC PRESS INC, COMPUTERS AND BIOMEDICAL RESEARCH; pp: 451-471;

Vol: 30

King Fahd University of Petroleum & Minerals

http://www.kfupm.edu.sa

Summary

This paper investigates the use of abductive-network machine learning for modeling

and predicting outcome parameters in terms of input parameters in medical survey

data. Here we consider modeling obesity as represented by the waist-to-hip ratio

(WHR) risk factor to investigate the influence of various parameters. The same

approach would be useful in predicting values of clinical parameters that are difficult

or expensive to measure from others that are more readily available. The AIM

abductive network machine learning tool was used to model the WHR from 13 other

health parameters. Survey data were collected for a randomly selected sample of 1100

persons aged 20 yr and over attending nine primary health care centers at Al-Khobar,

Saudi Arabia. Models were synthesized by training on a randomly selected set of 800

cases, using both continuous and categorical representations of the parameters, and

evaluated by predicting the WHR value for the remaining 300 cases. Models for

WHR as a continuous variable predict the actual values within an error of 7.5% at the

90% confidence limits. Categorical models predict the correct logical value of WHR

with an error in only 2 of the 300 evaluation cases. Analytical relationships derived

from simple categorical models explain global observations on the total survey

population to an accuracy as high as 99%. Simple continuous models represented as

analytical functions highlight global relationships and trends. Results confirm the

strong correlation between WHR and diastolic blood pressure, cholesterol level, and

family history of obesity. Compared to other statistical and neural network

approaches, AIM abductive networks provide faster and more automated model

synthesis. A review is given of other areas where the proposed modeling approach can

be useful in clinical practice. (C) 1997 Academic Press.

Copyright: King Fahd University of Petroleum & Minerals;http://www.kfupm.edu.sa

Page 2: © Modeling Obesity Using Abductive Networks Abdel-Aal, RE; Mangoud, AM ACADEMIC PRESS INC, COMPUTERS AND BIOMEDICAL RESEARCH; pp: 451-471; Vol: 30 King

1.2.3.4.5.6.7.8.9.10.11.12.13.14.15.16.17.18.19.20.21.22.23.24.25.26.27.28.29.30.31.32.33.34.35.36.37.38.39.40.41.42.43.

©

References:*ABT CORP, 1990, AIM US MANABDELAAL RE, 1994, ENERGY, V19, P739ABDELAAL RE, 1995, WEATHER FORECAST, V10, P310ABDELHALIM RE, 1993, SCAND J UROL NEPHROL, V27, P155ALLAIN CC, 1974, CLIN CHEM, V20, P470BARRON AR, 1984, SELF ORG METHODS MODBARRON RL, 1984, SELF ORG METHODS MODBAUMGARTNER RN, 1987, AM J EPIDEMIOL, V126, P614BRAY GA, 1988, W J MED, V149, P429BREIMAN L, 1984, CLASSIFICATION REGREBUCOLO G, 1973, CLIN CHEM, V19, P476CHARALAMBOUS C, 1992, IEE PROC-G, V139, P301DANIEL WW, 1974, BIOSTATISTICS FDN ANDENTONKELAAR I, 1990, NED TIJDSCHR GENEESK, V134, P1900DOYLE HR, 1995, METHOD INFORM MED, V34, P253DUCIMETIERE P, 1986, INT J OBESITY, V10, P229DUDA R, 1973, PATTERN RECOGNITIONFARLOW SJ, 1984, SELF ORG METHODS MODFOLSOM AR, 1990, AM J EPIDEMIOL, V131, P794HAFFNER SM, 1987, DIABETES, V36, P43HARTZ AJ, 1992, AM J CARDIOL, V70, P179IKEDA S, 1984, SELF ORG METHODS MODIVAKHNENKO AG, 1971, IEEE T SYST MAN CYB, V1, P364KAPLAN NM, 1989, ARCH INTERN MED, V149, P1514KENNEDY RL, 1990, CLIN SCI, V78, P24KEYS A, 1972, J CHRON DIS, V25, P329KISSEBAH AH, 1982, J CLIN ENDOCR METAB, V54, P254KNERR S, 1990, NEUROCOMPUTING ALGORKROTKIEWSKI M, 1983, J CLIN INVEST, V72, P1150LAPIDUS L, 1984, BRIT MED J, V289, P1261LAPIDUS L, 1988, INT J OBESITY, V12, P361LAPUERTA P, 1995, COMPUT BIOMED RES, V28, P38LAWS A, 1990, AM J PUBLIC HEALTH, V80, P1358LOWELL WE, 1994, J AM MED INFORM ASSN, V1, P459MALONE JM, 1984, SELF ORG METHODS MODMARSHAL SJ, P 2 INT C ART NEUR N, P200MOENS HJB, 1991, METHOD INFORM MED, V30, P187MONTGOMERY DC, 1985, INTRO LINEAR REGRESSMONTGOMERY GJ, P SPIE APPL ART NEUR, P56MYKKANEN L, 1993, DIABETOLOGIA, V36, P553OHLSON LO, 1985, DIABETES, V34, P1055OWENS A, P INT C NEUR NETW WA, P381

QUINLAN JR, 1987, INT J MAN MACH STUD, V27, P221Copyright: King Fahd University of Petroleum & Minerals;

http://www.kfupm.edu.sa

Page 3: © Modeling Obesity Using Abductive Networks Abdel-Aal, RE; Mangoud, AM ACADEMIC PRESS INC, COMPUTERS AND BIOMEDICAL RESEARCH; pp: 451-471; Vol: 30 King

44.45.46.47.48.49.50.51.52.53.54.55.

©

REVICKI DA, 1986, AM J PUBLIC HEALTH, V76, P992RUMELHART D, 1986, PARALLEL DISTRIBUTEDSCOTT DE, 1984, SELF ORG METHODS MODSOLER JT, 1988, J CLIN EPIDEMIOL, V41, P1075SZOLOVITS P, 1978, ARTIF INTELL, V11, P115VANSANT G, 1988, INT J OBESITY, V12, P133WADDEN TA, 1988, AM J CLIN NUTR, V47, P229WARNICK GR, 1978, J LIPID RES, V19, P65WEISS SM, 1991, COMPUTER SYSTEMS LEAWOOLERY LK, 1994, J AM MED INFORM ASSN, V1, P439WU YZ, 1993, RADIOLOGY, V187, P81ZHU K, P 2 INT C ART NEUR N, P205

For pre-prints please write to: [email protected]

Copyright: King Fahd University of Petroleum & Minerals;http://www.kfupm.edu.sa