A model is adequate, when it contains just enough parameters to describe the data, but not so many to describe the noise. Two methods are described to test this: The non-parametric runs test is quick and easy to perform, but not as powerful as the laborious t-test.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Testing Whether or Not a Model Fits the Data

  • Engelbert Buxbaum

摘要

A model is adequate, when it contains just enough parameters to describe the data, but not so many to describe the noise. Two methods are described to test this: The non-parametric runs test is quick and easy to perform, but not as powerful as the laborious t-test.