1. © GMDH-Based Feature Ranking And Selection For Improved Classification Of Medical Data Abdel-Aal, RE ACADEMIC PRESS INC ELSEVIER SCIENCE, JOURNAL OF BIOMEDICAL INFORMATICS; pp: 456-468; Vol: 38 King Fahd University of Petroleum & Minerals http://www.kfupm.edu.sa Summary Medical applications are often characterized by a large number of disease markers and a relatively small number of data records. We demonstrate that complete feature ranking followed by selection can lead to appreciable reductions in data dimensionality, with significant improvements in the implementation and performance of classifiers for medical diagnosis. We describe a novel approach for ranking all features according to their predictive quality using properties unique to learning algorithms based oil the group method of data handling (GMDH). An abductive network training algorithm is repeatedly used to select groups of optimum predictors from the feature set at gradually increasing levels of model complexity specified by the user. Groups selected earlier are better predictors. The process is then repeated to rank features within individual groups. The resulting full feature ranking can be used to determine the optimum feature subset by starting at the top of the list and progressively including more features until the classification error rate on all out- of-sample evaluation set starts to increase due to overfilling. The approach is demonstrated on two medical diagnosis datasets (breast cancer and heart disease) and comparisons are made with other feature ranking and selection methods. Receiver operating characteristics (ROC) analysis is used to compare classifier performance. At default model complexity, dimensionality reduction of 22 and 54%, Could be achieved for the breast cancer and heart disease data, respectively, leading to improvements in the overall classification performance. For both datasets, Copyright: King Fahd University of Petroleum & Minerals; http://www.kfupm.edu.sa

2. 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. © considerable dimensionality reduction introduced no significant reduction in the area under the ROC curve. GMDH-based feature selection results have also proved effective with neural network classifiers. (c) 2005 Elsevier Inc. All rights reserved. References: *ABT CORP, 1990, AIM US MAN ABBOTT DW, 1995, IEEE INT C SYST MAN, P1527 ABDELAAL RE, UNPUB COMPUT METHODS ABDELAAL RE, 1996, METHOD INFORM MED, V35, P265 ABDELAAL RE, 1997, COMPUT BIOMED RES, V30, P451 ABDULLA WH, 2003, INFORM SCIENCES, V156, P21, DOI 10.1016/S0020-0255(03)00162-2 AHA DW, 1996, COMP EVALUATION SEQU AHMAD A, 2005, PATTERN RECOGN LETT, V26, P43, DOI 10.1016/j.patrec.2004.08.015 BARRON AR, 1984, SELF ORG METHODS MOD, P87 BATTITI R, 1994, IEEE T NEURAL NETWOR, V5, P537 BLUM AL, 1997, ARTIF INTELL, V97, P245 BRILL FZ, 1992, IEEE T NEURAL NETWOR, V3, P324 CHEUNG RTF, 2001, CEREBROVASC DIS, V12, P1 DECHAZAL P, 2003, IEEE T BIO-MED ENG, V50, P686, DOI 10.1109/TBME.2003.812203 DELONG ER, 1988, BIOMETRICS, V44, P837 DETRANO R, 1989, AM J CARDIOL, V64, P304 DUCH W, 1999, INT JOINT C NEUR NET DUCH W, 2001, IEEE T NEURAL NETWOR, V12, P277 ECHAUZ J, 1994, P EL 94 INT C, P346 FARLOW SJ, 1984, SELF ORG METHODS MOD, P1 GARRETT D, 2003, IEEE T NEUR SYS REH, V11, P141, DOI 10.1109/TNSRE.2003.814441 GLETSOS M, 2003, IEEE T INF TECHNOL B, V7, P153, DOI 10.1109/TITB.2003.813793 HALL MA, 2000, P 17 INT C MACH LEAR HANLEY JA, 1983, RADIOLOGY, V148, P839 HASSANIEN AE, 2004, INFORMATICA-LITHUAN, V15, P23 HOFFMAN AJ, 1998, IEEE INT S IND EL, P663 KIRA K, 1992, P 9 INT C MACH LEARN, P249 KITTLER J, 1986, HDB PATTERN RECOGNIT, P59 KOHAVI R, 1997, ARTIF INTELL, V97, P273 KONDO T, 1999, P 38 SICE ANN C, P1181 KUPINSKI MA, 1997, INT C NEUR NETW, P2460 LEE WL, 2003, IEEE T MED IMAGING, V22, P382, DOI 10.1109/TMI.2003.809593 Copyright: King Fahd University of Petroleum & Minerals; http://www.kfupm.edu.sa

3. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. © LEWIN DI, 2000, IEEE INTELLIGENT SYS, V15, P2 MATSUI K, 1999, IEEE INT C IM PROC, P524 MCNITTGRAY MF, 1995, IEEE T MED IMAGING, V14, P537 MONTGOMERY GJ, 1990, P SPIE C APPL ART NE, P56 OPITZ DW, 1997, INT C NEUR NETW, P1400 OZDEMIR M, 2001, IEEE MOUNT WORKSH SO, P53 PACHEPSKY YA, 1999, SOIL SCI SOC AM J, V63, P1748 SWINIARSKI RW, 2003, PATTERN RECOGN LETT, V24, P833 VERIKAS A, 2002, PATTERN RECOGN LETT, V23, P1323 WOLBERG WH, 1990, P NATL ACAD SCI USA, V87, P9193 YU SX, 2000, IEEE SYS MAN CYBERN, P1912 ZARJAM P, 2003, INT S CIRC SYSTUNSP V-33-6 For pre-prints please write to: radwan@kfupm.edu.sa Copyright: King Fahd University of Petroleum & Minerals; http://www.kfupm.edu.sa