FE515 EViews Topics – Eviews Quick Start

Eviews – ECON SOFTWARE

Eviews Basics

Econometrics data

Eviews == Econometrics Eviews, written in C++.

Basic data types in econometrics are:

Cross-section data
Time-series data
Panel data

Objects

Eviews works in an “object” pattern, objects in Eviews are:

data
table
graph
equation
…

All objects stored in Workfile (wf1).

Actions can be done on object as :

copy
paste
save
rename
…

Eviews Workflow

Data preparation

Clean data.

Workfile Creation

2023-05-20-at-10.49.48

创建宏观经济案例的工作文件

宏观经济案例的工作文件视窗

Data import

导入外部xlsx数据文2：引导

导入外部xlsx数据文件3：检查

Object Generation (Genr)

Commonly used Eviews Objects

star	name_eng	role
4∗	Alpha	expressing variables that contain alphanumeric characters.
4∗	Coef	representing the parameters of an equation or system.
5∗	Equation	representing estimation, testing, or prediction of an equation.
5∗	Graph	representing graphical output results.
5∗	Group	representing a collection of sequential objects (including Series/Alpha, etc.).
4∗	Matrix	representing a matrix (a two-dimensional array).
4∗	Scalar	epresenting a scalar (i.e., a single value).
5∗	Series	representing a time series variable, with its elements required to be numeric.
5∗	Table	representing tabular output results.
4∗	Vector	column vector representing numeric values.

快速案例中的EViews对象一览

常用Eviews对象的菜单式创建操作

Exploratory Data Analysis (EDA)

with Graphs

id	purpose	name	characters
1	data comparison	(Circular Area Chart)	时期间比较->时期数较多->周期性时期
2	data comparison	折线图(Line Chart)	时期间比较->时期数较多->非周期性时期
3	data comparison	柱状图(Column Chart)	时期间比较->时期数较少->分组较少
4	data comparison	簇状-折线图(Line Chart)	时期间比较->时期数较少->分组较多
5	data comparison	不等宽柱状图(Variable Width Column Chart)	条目间比较->每个条目有两个变量->
6	data comparison	表格型嵌图(Table with Embedded Charts)	条目间比较->每个条目有一个变量->分组较多
7	data comparison	簇状-条形图(Bar Chart)	条目间比较->每个条目有一个变量->分组较少->条目较多
8	data comparison	簇状-柱状图(Column Chart)	条目间比较->每个条目有一个变量->分组较少->条目较少
1	data structure	饼形图(Pie Chart)	构成无时期变化->简单占比构成
2	data structure	瀑布图(Waterfall Chart)	构成无时期变化->累积构成或落差构成
3	data structure	百分比多层构成图(Components of Components)	构成无时期变化->多层次子类的百分比构成
4	data structure	堆栈百分比柱状图(Stacked 100% column chart)	构成有时期变化->时期较少->仅关心成分之间的相对差异
5	data structure	堆栈柱状图(Stacked Column Chart)	构成有时期变化->时期较少->仅关心成分之间的绝对差异
6	data structure	堆栈百分比面积图(Stacked 100% Area Chart)	Composition with periodic variations -> More periods -> Only interested in relative differences between components.
7	data structure	堆栈面积图(Stacked Area Chart)	Composition with periodic variations -> More periods -> Only interested in absolute differences between components.
1	data distribution	柱状直方图(Column Histogram)	Distribution of a single variable -> Few data points.
2	data distribution	线状直方图(Line Histogram)	Distribution of a single variable -> Numerous data points.
3	data distribution	散点图(Scatter Chart)	Distribution of two variables.
1	data correlation	3D面积图(3D Area Chart)	Distribution among three variables.
2	data correlation	散点图(Scatter Chart)	Relationship between two variables.
3	data correlation	泡泡图(Bubble chart)	Relationship between three variables.

Descriptive Statistics

构造主要变量的描述性统计分析表（三线表）

Graph Analysis

id	name_eng	mark
1	line & Symbol	Displaying a trend, including one sequence object, with numerous data points.
1	line & Symbol	Displaying a trend, including four sequence objects, with numerous data points.
2	Bar	Displaying group differences, horizontal bar chart, including one sequence object, with a small number of data groups.
2	Bar	Displaying group differences, vertical bar chart, including one sequence object, with a small number of data groups.
2	Bar	Displaying group differences, vertical bar chart, including several sequence objects, with a small number of data groups.
3	Spike	Displaying numerical magnitudes, including one sequence object, with a small number of data groups.
4	Area	Displaying coverage area, including one sequence object, with numerous data points.
4	Area	Displaying coverage area, including four sequence objects, with numerous data points.
5	Area Band	Displaying area differences, including two sequence objects, with numerous data points.
6	Mixed	Displaying diverse information, including multiple sequence objects, with numerous data points.
7	Dot Plot	Displaying numerical magnitudes, including one sequence object, with numerous data points.
8	Error Plot	Displaying numerical differences, including two sequence objects, with a small number of data points.
8	Error Plot	Displaying numerical differences, including three sequence objects, with a small number of data points.
9	High-Low	表现数值位置关系，包含2个序列对象（最高价、最低价），数据点较少
9	High-Low	表现数值位置关系，包含3个序列对象（开盘价、最高价、最低价），数据点较少
10	Scatter	表现变量关系，包含2个序列对象，数据点较多
10	Scatter	表现变量关系，包含多个序列对象，数据点较多
11	Bubble Plot	表现变量关系，包含3个序列对象，数据点较多
12	XY Line	表现数据变化路径，包含2个序列对象，数据点较多
13	XY Area	表现2D覆盖面积，包含个序列对象，数据点较多
14	XY Bar	表现分组差异，垂直条形，包含若干个序列对象，数据分组较少
15	Pie	表现数据构成关系，包含1个序列对象，数据分组较少
16	Distribution	表现数据分布密度，直方图，包含1个序列对象，数据点较多（自动分组）
17	Quantile-Quantile	表现数据分布密度，正态QQ图，包含1个序列对象，数据点较多
18	Boxplot	表现数据分布形态，包含若干序列对象，数据点较多
19	Seasonal Graph	表现数据循环关系，包含1个序列对象，数据点较多，分组较少（在同一月份不同年份上的变化）
19	Seasonal Graph	表现数据循环关系，包含1个序列对象，数据点较多，分组较少（在同一年份不同月份上的变化）

Line-Symbol

主要变量的简单多线图（Line-Symbol）绘制

主要变量的个性化多线图（Line Symbol）绘制

Linear Regression Analysis

主回归模型Eviews操作

Command

$通过命令视窗提取回归方程的判定系数$R^2$和调整判定系数$\bar{R}^2$$

$通过命令视窗计算得到理论$t$值$t_{1-\alpha/2}(f)=t_{0.975}(125)$$

Program (Batch Commands)

EViews程序文件（.prg）的创建操作

Command Syntax (program syntax)

Syntax:


x
action(action_opt) object_name.view_or_proc(options_list) arg_list

action:
- show : show result
- do : run a program
- freeze : generate table or graph
- print : print the result
action_opt : options for action
object_name: Object name of workfile
view_or_proc : View and Procedure Commands
- ser1,sheet, ser1.stats , ..
- ser1.line,ser1.bar,ser1.hist, …
- ser1.freq,ser1.correl,ser1.lrval,ser1.uroot,ser1.vratio 2 4 8 ,ser1.bdstest , …
- …
options_list : View and procedure Commands Options
arg_list : arguments
- gdp c investment pce government $gdp=\hat{\beta_1}+\hat{\beta_2}investment+\hat{\beta_3}pce+\hat{\beta_4}government+e_i$

Commonly used commands and functions

2023-05-20-at-11.25.31

2023-05-20-at-11.25.44

2023-05-20-at-11.26.03

A univariate linear regression model (SLR)

OLS

Essense of Regression:

\begin{matrix} (1) & Y_{i} - - - - > Y | X_{i} \end{matrix}

Ordinary Least Squares (OLS): Minimizes the sum of squared residuals between observed values and the predicted values of a regression model to estimate population parameters.
Maximum Likelihood Estimation (MLE): Assumes that the observed data is generated from a probability distribution and estimates population parameters by maximizing the likelihood function of the observed data.
Bayesian Estimation: Incorporates prior knowledge and beliefs about the parameters by specifying prior distributions, and updates them using observed data to obtain the posterior distribution of the parameters.
Method of Moments (MOM): Matches the sample moments (such as means and variances) to their population counterparts to estimate population parameters.

OLS

\begin{aligned} (2.2) & Y_{i} & = β_{1} + β_{2} X_{i} + u_{i} & (PRM) \\ (2) & u_{i} & \sim N (0, σ^{2}) \\ (2.3) & Y_{i} & = {\hat{β}}_{1} + {\hat{β}}_{2} X_{i} + e_{i} & (SRM) \end{aligned}

\begin{aligned} (3) & M i n (Q) & = \sum (Y_{i} - {\hat{Y}}_{i})^{2} \\ (4) & = \sum {(Y_{i} - ({\hat{β}}_{1} + {\hat{β}}_{2} X_{i}))}^{2} \\ (5) & = f ({\hat{β}}_{1}, {\hat{β}}_{2}) \end{aligned}

\begin{aligned} (2.4) & \begin{aligned} {\hat{β}}_{2} & = \frac{n \sum X_{i} Y_{i} - \sum X_{i} \sum Y_{i}}{n \sum X_{i}^{2} - {(\sum X_{i})}^{2}} \\ \hat{β_{1}} & = \frac{n \sum X_{i}^{2} Y_{i} - \sum X_{i} \sum X_{i} Y_{i}}{n \sum X_{i}^{2} - {(\sum X_{i})}^{2}} \end{aligned} & (FF solution) \end{aligned}

\begin{aligned} (2.5) & \begin{aligned} {\hat{β}}_{2} & = \frac{\sum x_{i} y_{i}}{\sum x_{i}^{2}} \\ \hat{β_{1}} & = {\bar{Y}}_{i} - {\hat{β}}_{2} {\bar{X}}_{i} \end{aligned} & (ff solution) \end{aligned}

\begin{aligned} (2.6) & {\hat{Y}}_{i} & = {\hat{β}}_{1} + {\hat{β}}_{2} X_{i} \\ (2.7) & e_{i} & = Y_{i} - {\hat{Y}}_{i} \\ (2.8) & {\hat{σ}}^{2} & = \frac{\sum e_{i}^{2}}{n - 1} = \frac{\sum (Y_{i} - {\hat{Y}}_{i})^{2}}{n - 1} \end{aligned}

T-statistic

The t-statistic is used to test the hypothesis that the coefficient is different from zero ( null hypo).

for a PRM:

\begin{aligned} (6) & Y_{i} & = β_{1} + β_{2} X_{i} + u_{i} \end{aligned}

Bigger t-statistic, higher the significance. lower p-val higher significance.

F-value

This used to test the overall significance of the model.

Bigger F-value , higher the significance. lower p-val higher significance.

How to read Eviews SLM report

看懂EViews回归分析报告

\begin{aligned} (7) & Y_{i} & = β_{1} + β_{2} X_{i} + ϵ_{i} \end{aligned}

Variable : Names of Var
- C constant : intercept
- INCOME : X
Coefficient: Coefficients
- $^β1$
- $^β2$
Std. Error : of the coeff
- $S_{\hatβ_1}$
- $S_{\hatβ_2}$
t-Statistic :
- $t∗\hat β1$
- $t∗\hat β2$
Prob.:
- $p(\alpha/2, t^{\ast}_{\hat{\beta}_1}, f)$
- $p(\alpha/2, t^{\ast}_{\hat{\beta}_2}, f)$

生成回归方程对象

主回归模型Eviews操作