The effects of the underlying disease and serum albumin on GFR prediction using the Adaptive Neuro Fuzzy Inference System (ANFIS)

Document Type : Articles

Authors

Abstract

Introduction: Kidney disease is a major public health challenge worldwide. Epidemiologic data suggest a significant relationship between underlying diseases and decrease in Glomerular Filtration Rate (GFR). Clinical studies and laboratory research have shown that the mentioned parameter is effective in development and progression of the renal disease per se. In this study, we used learning-based system based on the neural network concepts. Method: To predict GFR and propose an intelligent method with few errors (about 3%), we need to prognosticate the course and severity of the kidney disease in patients with chronic kidney disease using limited data and information. Adaptive neuro fuzzy inference system (ANFIS) used in the present study is based on the model proposed by Jang, and all laboratory (creatinine, calcium, phosphorus, albumin) and underlying disease caused by chronic kidney disease ( CKD ) were reviewed.Results: It has been shown that the rate of GFR decreases in patients with diabetes and glomerulopathy was faster than other causes. Furthermore, serum albumin level less than 4.5gr/dl with diabetes was also associated with higher risk of rapid GFR loss.Conclusion:Therefore, it seems that this modeling of fuzzy variables with error less than 3.5% in some cases and create a fuzzy inference system model that presents the complex relationships between the laboratory input variables and GFR as simple linear models.Keywords: GFR, ANFIS, Underlying kidney disease, Albumin, CKD

  1. Mahdavi-Mazdeh M. Why do we need chronic kidney
  2. disease screening and which way to go? Iranian journal
  3. of kidney diseases. [Review]. 2010;4(4):275-81.
  4. Song Q, Kasabov N. TWNFI--a transductive
  5. neuro-fuzzy inference system with weighted data
  6. normalization for personalized modeling. Neural
  7. networks : the official journal of the International
  8. Neural Network Society. [Comparative Study Research
  9. Support, Non-U.S. Gov’t]. 2006;19(10):1591-6.
  10. Gaspari F, Ferrari S, Stucchi N, Centemeri E, Carrara F,
  11. Pellegrino M, et al. Performance of different prediction
  12. equations for estimating renal function in kidney
  13. transplantation. American journal of transplantation
  14. : official journal of the American Society of
  15. Transplantation and the American Society of Transplant
  16. Surgeons. 2004;4(11):1826-35.
  17. Brier ME, Ray PC, Klein JB. Prediction of delayed
  18. renal allograft function using an artificial neural
  19. network. Nephrology, dialysis, transplantation :
  20. official publication of the European Dialysis and
  21. Transplant Association - European Renal Association.
  22. ;18(12):2655-9.
  23. Polat K, Güneş S. An expert system approach based
  24. on principal component analysis and adaptive neuro-
  25. fuzzy inference system to diagnosis of diabetes disease.
  26. Digital Signal Processing. 2007;17(4):702-10.
  27. Polat K, Güneş S. Automatic determination of diseases
  28. related to lymph system from lymphography data using
  29. principles component analysis (PCA), fuzzy weighting
  30. GFR Prediction by Fuzzy Expert System
  31. Figure 6. The impact of underlying disease on GFR in the
  32. control group
  33. J Health Man & Info. 2014;1(3):50
  34. pre-processing and ANFIS. Expert Systems with
  35. Applications. 2007;33(3):636-41.
  36. Sengur A. An expert system based on principal
  37. component analysis, artificial immune system and
  38. fuzzy k-NN for diagnosis of valvular heart diseases.
  39. Computers in biology and medicine. 2008;38(3):329-38.
  40. Jang JSR. Neuro-Fuzzy Modeling: Architectures,
  41. Analyses and Applications. Berkeley CA 94720:
  42. University of California; 1992.
  43. Jang J-SR, Sun C-T, Mizutani E. Neuro-fuzzy and
  44. soft computing-a computational approach to learning
  45. and machine intelligence [Book Review]. Automatic
  46. Control, IEEE Transactions on. 1997;42(10):1482-4.
  47. Sivasankar E, Rajesh RS. Knowledge Discovery
  48. in Medical Datasets Using a Fuzzy Logic rule
  49. based Classifier. Electronic Computer Technology
  50. Internationalconfrence, IEEE. 2010:208 - 13.
  51. Han M, Snow PB, Brandt JM, Partin AW. Evaluation of
  52. artificial neural networks for the prediction of pathologic
  53. stage in prostate carcinoma. Cancer. [Evaluation
  54. Studies]. 2001;91(8 Suppl):1661-6.
  55. MA TM. Local and Personalised Modelling for
  56. Renal Medical Decision Support Systems, Auckland
  57. University. 1996 ed. Neural network design. Boston,
  58. USA: PWS Pub Co.; 2012.