عناصر مشابهة
Lower and Upper Probabilities for Multiple Comparisons
| المصدر: | مجلة العلوم والدراسات الإنسانية |
|---|---|
| الناشر: |
جامعة بنغازي - كلية الآداب والعلوم بالمرج
|
| المؤلف الرئيسي: | |
| مؤلفين آخرين: | |
| المجلد/العدد: | ع14 |
| محكمة: | نعم |
| الدولة: | ليبيا |
| التاريخ الميلادي: | 2016 |
| الصفحات: | 1 - 21 |
| DOI: | 10.37376/1571-000-014-005 |
| ISSN: | 2312-4962 |
| رقم MD: | 763215 |
| نوع المحتوى: | بحوث ومقالات |
| اللغة: | English |
| قواعد المعلومات: | HumanIndex |
| مواضيع: | |
| رابط المحتوى: |
|
| المستخلص: | In recent years, nonparametric predictive inference (NPI) has been developed as a frequentist statistical framework which requires only few modelling assumptions and considers one or more future observations instead of a population. NPI uses lower and upper probabilities for uncertainty quantication, and has strong consistency properties within theory of interval probability. As NPI does not aim at inference for an entire population but instead explicitly considers one or more future observations, this provides an attractive alternative to standard methods of frequentist statistical inference. This work presents NPI for multiple comparisons and subset selection for ordinal categorical data. For events which are of most practical interest, closed-form formulae for the NPI lower and upper probabilities are derived. Some special cases are discussed and illustrated with examples. |
|---|
