多元回归分析之序列相关.pptVIP

Intermediate Econometrics, Yan Shen Chapter Outline本章大纲 Properties of OLS with Serially Correlated errors 误差序列相关时OLS的性质 Testing for Serial Correlation 检验序列相关 Correcting for Serial Correlation with Strictly Exogenous Regressors 当自变量为严格外生时校正序列相关 Differencing and Serial Correlation 差分和序列相关 Heteroskedasticity in Time Series Regression 时间序列回归中的异方差性 Lecture Outline讲义大纲 What is serial correlation 什么是序列相关 Basic introduction of time series analysis 时间序列分析的基本介绍 Properties of OLS with Serially Correlated Errors 误差序列相关时OLS的性质 Testing for Serial Correlation 检验序列相关 What is serial correlation什么是序列相关 Serial correlation happens when the covariances of the error terms are not zero, that is, for some individuals i and m, 当误差项协方差不为零时，序列相关就出现了。即，对某些观察值i和m, cov(ui,um)?=0. Even though the problem of serial correlation can also happen to cross-section data when the data are ordered in a specific way, it is a frequent one when using time series due to inertia in the system. 分析截面数据时，如果我们把数据按特定方式排序，序列相关的问题也可能发生，然而由于系统产存在时间上的惰性，它在时间序列分析中更为常见。 For this reason it is often called autocorrelation. 因为这个原因，它常常被称作是自相关。 Basic Regression Analysis with Time Series Data时间序列数据的基本回归分析 We focus on discussing the Gauss-Markov assumptions for time series applications. 我们集中讨论时间序列版的高斯-马尔可夫假定。 The Nature of Time Series Data 时间序列数据的本质 A time series data set is a sequence of random variables indexed by time. 时间序列数据是以时间为指标的一个随机变量序列。 Time series data set comes with a temporal ordering. 时间序列数据集伴随着一个时间上的排序。 Basic Regression Analysis with Time Series Data时间序列数据的基本回归分析 Example: a static model 例：一个静态模型 A dynamic model一个动态模型 Time Series Data: Finite Sample Properties of OLS Under Classical Assumptions时间序列数据：在经典假定下OLS的有限样本性质 Unbiasedness of OLS OLS的无偏性 Assumption TS.1: Linear in parameters 假定 TS.1: 模型对于参数呈线性关系 Assumption TS.2: Zero conditional mean 假定 TS.2: 零条件期望 Assumption TS.3: No perfect collinearity 假定 TS.3：没有完全共线性 The