Page 84 - Basic Statistics

79

5. LINEAR REGRESSION MODEL

Linear regression is a statistical analysis that models the relationship

several variables by linear equations explicit form of relationship. Explicit form
of the linear equation is a linear equation that puts a single variable is on one

side of an equation.

Explicit variable in the model is a random variable, and the most likely to

have behavior that depends on other variables. Variables which is the main

concern is expressed as a dependent variable (response), with the symbol Y. As
an example for these variables, can be a death caused by a disease, the level of

prices according to market conditions, and the learning achievement of a

teaching method.

Other variables in a model of linear equations are variables that might

provide information about the behavior of dependent variables Y. These
variables are placed as a predictor or independent variables in the model of

linear equations. These variables are variables that are known fixed (not

random), hereinafter referred to as independent variables, with the symbol X.

In general, this linear regression modeling aims to present how the

average value of dependent variable "E(Y)" changes according to the change of
each independent variable. It is assumed that the variance of Y is unaffected by

changes in each independent variable. Furthermore, the linear regression

equation is expressed as the seat of the expectation value of Y at each X value

which is fixed. This expectation values have identical distribution and variance

are equal.

5.1 THE SIMPLE LINEAR REGRESSION MODEL

5.1.1 MODEL AND ESTIMATION COEFFICIENTS

Simple linear regression model involves only one independent variable

X. This model states constantly change the average value of the response

~~* CHAPTER 5 LINEAR REGRESSION MODEL *~~