原创力文档- 1、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。。
- 2、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。
- 3、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)。
- 4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
- 5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
- 6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
- 7、VIP文档为合作方或网友上传，每下载1次， 网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档
查看更多
1 Time Series Model and Forecasting 11 Introduction - iSites1时间序列模型与11个引进isites预测
PAGE
PAGE 9
Econometrics 1123: Section 9 Handout
Eunice Han
ehan@fas.harvard.edu
November 18th, 2010
1 Time Series Model and Forecasting
1.1 Introduction
There is only one observation at one time. When we analyze time series data, we tend to say we are interested in “impulse(shock)-response” relationship. Upmost purpose in time series study is often to make a good “forecast”.
?Examples:
-Impact of monetary policy on aggregate demand
-Impact of exchange rate on export
-Impact of rise in oil price on stock market index
1.2 Stationarity
Time series Yt is stationary if its distribution does not change over time. It implies that history is relevant for forecasting.
?Time series Yt is stationary if
(1) E(Yt) is constant over time
(2) Var(Yt) is constant over time
(3) Cov(Yt, Yt-j) only depends on j, not on t
Q. Why do we need “stationary” assumption? If each random variable has different expectation and variance, and , where t=1, 2, …,T. Can we estimate and ?
No. In time series data, we have only one observation for each random variable. So we cannot get mean and variance of each random variable. And this is why we need stationarity assumption.
1.3 Unconditional vs Conditional Expectation
? Unconditional Expectation: Time series Yt has unconditional mean , if it is stationary.
? Conditional Expectation: We can make a prediction using up-to-date information of Yt . (Ex.)
Figure 1: Conditional and unconditional expectation
1.4 Autoregressive Models: AR(p)
1. The autoregressive model of order p, called AR(p) model, represents Yt as a linear function of p of its lagged values:
?We need the assumption of E(ut | Yt-1,Yt-2,… Yt-p)=0. The error terms are serially uncorrelated.
?The coefficient does not have a causal interpretation.
?We determine p using information criterion such as Bayesian Information Criterion (BIC) or Akaike Information criterion (AIC).
?Yt has to be stationary; otherwise our “normal” regression theory breaks down.
Now, suppose we have AR(1) mode
文档评论（0）