دوره 7، شماره 2 - ( 4-1398 )
جلد 7 شماره 2 صفحات 60-49 |
برگشت به فهرست نسخه ها
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Behnampour A, Biglarian A, Bakhshi E. Application of fuzzy logistic regression in modeling the severity of autism spectrum disorder. Jorjani Biomed J 2019; 7 (2) :49-60
URL: http://goums.ac.ir/jorjanijournal/article-1-657-fa.html
URL: http://goums.ac.ir/jorjanijournal/article-1-657-fa.html
Application of fuzzy logistic regression in modeling the severity of autism spectrum disorder. فصلنامه علمی پژوهشی زیست پزشکی جرجانی. 1398; 7 (2) :49-60
URL: http://goums.ac.ir/jorjanijournal/article-1-657-fa.html
چکیده: (6465 مشاهده)
Background and objectives: Autism spectrum disorder (ASD) is a childhood neurodevelopmental disorder and according to DSM-5 classification, its severity includes three levels: requiring support, requiring substantial support, and requiring very substantial support. This classification is unclear from a possible perspective and from a fuzzy point of view; it has a degree of uncertainty. The purpose of this study is to predict the severity of autism disorder by fuzzy logistic regression.
Methods: In this cross-sectional study, 22 children with ASD which referred to the rehabilitation centers of Gorgan in 2017 were used as a research sample. Therapist's viewpoint about the severity of the disorder that is measured by linguistic terms (low, moderate, high) was considered as fuzzy output variable. In addition, to determine the prediction model for the severity of autism, a fuzzy logistic regression model was used. In this sense parameters were estimated by least square estimations (LSE) and least absolute deviations (LAD) methods and then the two methods were compared using goodness-of-fit index.
Results: The age of children varied from 6 to 17 years old with mean of 10.44± 3.33 years. Also, the goodness-of-fit index for the model that was estimated by the LAD method was 0.0634, and this value was less than the LSE method (0.1255). The estimated model by the LAD indicates that with the constant of the values of other variables, with each unit increase in the variables of age, male gender, raw score of stereotypical movements, communication and social interaction subscales, possibilistic odds of severity of autism disorder varied about 0.67 (decrease), 0.362 (decrease), 0.098 (increase), 0.019 (increase) and 0.097 (increase) respectively.
Conclusion: The LAD method was better than LSE in parameter estimation. So, the estimated model by this method can be used to predict the severity of autism disorder for new patients who referred to rehabilitation centers and according to predicted severity of the disorder, proper treatments for children can be initiated.
Methods: In this cross-sectional study, 22 children with ASD which referred to the rehabilitation centers of Gorgan in 2017 were used as a research sample. Therapist's viewpoint about the severity of the disorder that is measured by linguistic terms (low, moderate, high) was considered as fuzzy output variable. In addition, to determine the prediction model for the severity of autism, a fuzzy logistic regression model was used. In this sense parameters were estimated by least square estimations (LSE) and least absolute deviations (LAD) methods and then the two methods were compared using goodness-of-fit index.
Results: The age of children varied from 6 to 17 years old with mean of 10.44± 3.33 years. Also, the goodness-of-fit index for the model that was estimated by the LAD method was 0.0634, and this value was less than the LSE method (0.1255). The estimated model by the LAD indicates that with the constant of the values of other variables, with each unit increase in the variables of age, male gender, raw score of stereotypical movements, communication and social interaction subscales, possibilistic odds of severity of autism disorder varied about 0.67 (decrease), 0.362 (decrease), 0.098 (increase), 0.019 (increase) and 0.097 (increase) respectively.
Conclusion: The LAD method was better than LSE in parameter estimation. So, the estimated model by this method can be used to predict the severity of autism disorder for new patients who referred to rehabilitation centers and according to predicted severity of the disorder, proper treatments for children can be initiated.
نوع مقاله: تحقیقی |
موضوع مقاله:
آمار زیستی
دریافت: 1397/11/18 | پذیرش: 1398/1/21 | انتشار: 1398/4/10
دریافت: 1397/11/18 | پذیرش: 1398/1/21 | انتشار: 1398/4/10
| بازنشر اطلاعات | |
 |
این مقاله تحت شرایط Creative Commons Attribution-NonCommercial 4.0 International License قابل بازنشر است. |
