Abstract:

A prediction model of interval grey number based on kernel and grey degree preliminaryly solves the problem of prediction for interval grey number, while still has some model deficiencies which need to be improved in error analysis and the method of determining the predicted value of grey degree of interval grey number. These model deficiencies are analyzed based on the definition and characteristics of compound grey number which are proposed. Furthermore, a prediction model for grey degree is built instead of the original deficient method, thus the original prediction model of interval grey number is improved and perfected. The improved model can excavate both potential information and developing trend of interval grey number list from the aspects of both kernel and grey degree, thus the original model deficiencies are overcome, and error analysis and precision test are supported. Example analysis demonstrates the applicability of the improved model.