Fuzzy simple linear regression using Gaussian membership functions minimization problem
Under the additional assumption that the errors are normally distributed, the Ordinary Least Squares (OLS) method is the maximum likelihood estimator. In this paper, we propose, for a simple regression, an estimation method alternative to the OLS method based on a so-called Gaussian membership funct...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Ayandegan Institute of Higher Education,
2022-10-01
|
| Series: | Journal of Fuzzy Extension and Applications |
| Subjects: | |
| Online Access: | https://www.journal-fea.com/article_155107_24583b21a95541723ceb8e7bc158641e.pdf |
| Summary: | Under the additional assumption that the errors are normally distributed, the Ordinary Least Squares (OLS) method is the maximum likelihood estimator. In this paper, we propose, for a simple regression, an estimation method alternative to the OLS method based on a so-called Gaussian membership function, one that checks the validity of the verbal explanation suggested by the observer. The fuzzy estimation approach demonstrated here is based on a suitable framework for a natural behavior observed in nature. An application based on a group of MENA countries in 2015 is presented to estimate the employment poverty relationship.Under the additional assumption that the errors are normally distributed, the ordinary least squares method is the maximum likelihood estimator. In this paper, we propose, for a simple regression, an estimation method alternative to the ordinary least squares method based on a so-called Gaussian membership function, one that checks the validity of the verbal explanation suggested by the observer. The fuzzy estimation approach demonstrated here is based on a suitable framework for a natural behavior observed in nature. An application based on a group of MENA countries in 2015 is presented to estimate the Employment Poverty relationship. |
|---|---|
| ISSN: | 2783-1442 2717-3453 |