Scribd
Upload
15 views2 pages

Ria Stats Regression Analysiss

Regression analysis is a statistical method for determining the relationship between variables, introduced by Francis Galton in 1886, commonly used in various fields such as finance and psychology. Simple linear regression involves two variables, where one is independent and the other is dependent, and is represented by the equation Yi = α + βXi + εi. Key assumptions for this model include linearity, independence of errors, normality of error distribution, and homoscedasticity, with properties of regression coefficients highlighting their correlation and sign consistency.

Uploaded by

riavprakash
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
15 views2 pages

Ria Stats Regression Analysiss

Regression analysis is a statistical method for determining the relationship between variables, introduced by Francis Galton in 1886, commonly used in various fields such as finance and psychology. Simple linear regression involves two variables, where one is independent and the other is dependent, and is represented by the equation Yi = α + βXi + εi. Key assumptions for this model include linearity, independence of errors, normality of error distribution, and homoscedasticity, with properties of regression coefficients highlighting their correlation and sign consistency.

Uploaded by

riavprakash
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

1.

REGRESSION ANALYSIS
DEFINITION AND TYPES :

Regression analysis deals with determining the relation between two or more variables. In
regression analysis, we estimate the values of parameters involved in the regression equation.
The idea of regression was introduced by Francis Galton (1822-1911) in his paper published in
1886.
For example, income-expenditure data of the form (Xi, Yi), i = 1,2,…,n, where Xi and Yi
represents income and expenditure respectively of n-th household.
The aim will be to develop a regression model of the form ‘Y = a+bX’. Once we develop the
regression model between income and expenditure, for a given level of income Xi, expenditure
can be predicted. The regression equation will also help to understand how expenditure is
related with income. Application of regression exists in large numbers or amounts. It is used in
finance, economics, psychology, and agriculture science to name a few. Regression can be of
various types; I’ll be talking about simple linear regression.

In the example of income-expenditure data (Xi, Yi), one can expect a positive value of β in the
developed regression model. Similarly, one can expect a negative slope for the regression
model where Xi and Yi represent the price and sale of certain products.

2. SIMPLE LINEAR REGRESSION


Linear regression involving only two variables is called simple linear regression.
Let us consider two variables as ‘x’ and ‘y’. Here ‘x’ represents an independent variable or
explanatory variable and ‘y’ represents a dependent variable or response variable. Dependent
variable must be a ratio variable, whereas an independent variable can be a ratio or categorical
variable.

A linear pattern can easily be identified in the data by plotting the scatter diagram
For example : Let population regression model is given by: Yi = α + βXi + εi , i = 1,2,…,n .

Where Xi and Yi are independent (explanatory) and dependent (response) variables


respectively. α is intercept and β is slope of the regression line. εi represents random error or
noise in Y-value for i-th observation. The regression equation can be split into two parts. First
part (Yi = α + βXi) is a representation of a straight line, whereas the second part εi indicates the
error term. In the first part of the regression model, intercept α is the mean value of a dependent
variable for an independent variable equal to zero. Slope β is the rate of change of the
dependent variable. Equivalently, β is the amount of change in dependent variable (Y) for one
unit of change in independent variable (X).

A positive value of slope indicates a positive correlation between the variables

It should be noted here that the independent variables (Xi’s) are fixed, whereas the dependent
variable (Yi) is assumed to be random. Corresponding to one given value of Xi , we can have
multiple values of Yi. Therefore, Yi has a probability distribution. The error term εi is also a
random variable and with regard to a given value of Xi , we can have different errors.
3. Assumptions of a simple linear regression model
The important assumptions of a simple linear regression model are given below:
(i) Linearity: The first assumption of the model is that there exists a linear relationship between
the variables
(ii) Independence: The second assumption states that residuals or error terms are independent
to each other
(iii) Normality: The error terms εi follows a normal distribution with mean zero and a constant
variance
(iv) Homoscedasticity: The error terms have equal variances for each value of X. This property
of equal variances is also called homoscedasticity. Before we develop a regression model, we
should check that none of the assumptions is violated. Any violation of the above assumptions
may impact the accuracy of the model.

4. Properties of regression coefficients


(i) Karl Pearson coefficient of correlation can be viewed as the geometric mean of the two
regression coefficients.
(ii) It should be noted here that the two regression coefficients must be of the same sign. That
is, both of them should be either positive or negative in sign.
(iii) If one regression coefficient is greater than one, then the other must be less than one.
(iv) The regression coefficients are independent of change of origin but not of scale.

You might also like