컨텐츠 시작
학술대회/행사
초록검색
| 제출번호(No.) | 0540 |
|---|---|
| 분류(Section) | Contributed Talk |
| 분과(Session) | Probability / Stochastic Process / Statistics (SS-12) |
| 영문제목 (Title(Eng.)) |
Fuzzy logistic regression for fuzzy categorical response data |
| 저자(Author(s)) |
Mahshid Namdari1, Mahmoud Taheri2, Alireza Abadi1, Farhad Niaghi1, Seung Hoe Choi3 Shahid Beheshti University of Medical Sciences, Tehran, Iran1, University of Tehran, Tehran, Iran2, Korea Aerospace University3 |
| 초록본문(Abstract) | Statistical logistic regression is used for modeling a response variable with crisp (exact/non fuzzy) categories based on a set of explanatory variables. But, in real world, there are many situations in which, due to lack of suitable instruments or well defined criteria, the borders between the response categories are vague and we are encountered with non-precise observations. In these situations, the probabilistic assumptions of the ordinary logistic regression model are not fulfilled and fuzzy logistic regression could be an alternative choice. In this paper, a fuzzy logistic regression is investigated, which can be used in cases where the explanatory variables are crisp observations but the values of the response variable is non-precise. We attempted to improve the model by bringing the membership functions of the estimated responses as close as possible to those of the corresponding observed values. A numerical example in a real clinical study about the relationship between the severity of glioma tumors and a set of risk factors is investigated by the proposed regression model. Finally, the obtained model is evaluated by the means of a goodness-of-fit index. |
| 분류기호 (MSC number(s)) |
62J86 |
| 키워드(Keyword(s)) | fuzziness, linear inference, regression |
| 강연 형태 (Language of Session (Talk)) |
English |