Scribd
Upload
2 views

8.Linear regression-correlation

The document discusses simple linear regression, including its model, equation, and evaluation methods. It covers key concepts such as correlation coefficients and the coefficient of determination, along with assumptions and precautions for accurate analysis. Examples are provided to illustrate the application of these statistical techniques in predicting relationships between variables.

Uploaded by

maittt.22ba13211
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
2 views

8.Linear regression-correlation

The document discusses simple linear regression, including its model, equation, and evaluation methods. It covers key concepts such as correlation coefficients and the coefficient of determination, along with assumptions and precautions for accurate analysis. Examples are provided to illustrate the application of these statistical techniques in predicting relationships between variables.

Uploaded by

maittt.22ba13211
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 32

Nguyen Thi Van Anh, PhD

Department of Pharmacological, Medical, Agronomical Biotechnology


Department
Contents
• The simple linear regression model
• The simple linear regression equation
• Evaluating the simple linear regression equation
• The correlation coefficient
• The coefficient of determination
• Some precautions
1. INTRODUCTION
• Regression
to predict or estimate the value of one variable corresponding to
a given value of another variable

• Correlation
the strength of the relationship between variables

2 techniques to explore Relationship between variables


2. Simple linear regression
Objectives:
2. Simple linear regression model

Assumptions Underlying Simple Linear Regression

1. Values of the independent variable X are said to be


“fixed.” (preselected by investigators)
2. The variable X is measured without error.
3. For each value of X there is a subpopulation of Y values.
These subpopulations must be normally distributed.
Assumptions Underlying Simple Linear Regression
Linear regression model:
Method of least square

General equation:
The Least-Squares Criterion
The sum of the squared ver0cal devia0ons of the
observed data points (yi) from the least-squares line
is smaller than the sum of the squared ver0cal
devia0ons of the data points from any other line.
Measure of deviation
yi = values observed; : mean; : values from regression line
Total sum of square Regression sum of Error sum of
(SST) square (SSR) square (SSE)
Measure of varia/on
Obtaining the least square line:
From sample data, carried out by hands using equations:
Example:
A substance used in biological and medical research is shipped
by airfreight to users in cartons of 1000 ampules. The data
below, involving 10 shipments, were collected on the number of
times the carton was transferred from 1 aircraft to another over
the shipment route (X) and the number of ampules found to be
broken upon arrival (Y)
3. Calcula/ng a correla/on coefficient
THE CORRELATION COEFFICIENT

r < 0: negative linear relationship


r > 0: positive linear relationship
4. The Coefficient of DeterminaAon

Ratio between explained variation and total variation

r2 = 1: perfect linear relationship


0 < r2 < 1: weaker linear relationship
r2 = 0: no linear relationship
Example:
Example: Compute correlation coefficient r?
Example:
5. Evaluating the regression equation

Test H0: β1 = 0

Equation: little/no value for prediction/estimation

(1) The relationship is linear, sufficient strength to use the sample


regression equation to predict and estimate Y from X
(2) There is a good fit of data to a linear model, but a curvilinear model
might predict even better.
Is there any linear relationship between X and Y?
Regression equation obtained from sample:
SSE =
Example:

Y = 98.25 + 0.1098 X

α = 0.05
= 0.03297

= SSE =

> 2.1199
• Interpreting the Results
6. Some precautions
• The assumptions underlying regression and correlation analysis
should be reviewed carefully before the data are collected.

• No matter how strong is the indication of a relationship between 2


variables, it should not be interpreted as one of cause and effect.

• The sample regression equation should not be used to predict or


estimate outside the range of values of the independent variable
represented in the sample

You might also like