- 1、本文档共28页,可阅读全部内容。
- 2、原创力文档(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。
- 3、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载。
- 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
查看更多
EVIEWS 极大似然法.pdf
MAXIMUM
LIKELIHOOD
ESTIMATION
B. Essama-Nssah
Poverty Reduction Group (PRMPR)
The World Bank
Washington, DC
Module 1
Outline
Principle
Implementation
The Logl Object
Simple Regression Model
Hands-On Exercise
2
Principle
Consider a sample of n observations on two variables (y , x).
i i
Suppose the data are generated by a stochastic process of
the following form.
y β +β x +ε
i1 2 i i
Where the ε are independent random normal variates with
i
mean zero and standard deviation σ.
To estimate the underlying parameters by the ML method, we
need to find values of these parameters that would make the
available observations most likely. Indeed, different
population parameters are bound to generate different
samples.
3
Principle
The parameters of interest here are β , β ,
1 2
and the variance σ2.
To estimate these, we need to find
estimators (i.e. formulae) that maximize
the likelihood (or log likelihood) function
based on sample observations.
For implementation in EViews, it is more
convenient to consider the log likelihood
function as the sum of the log likelihood
contributions for each observation i.
4
文档评论(0)