BIVariate and Multivariate Regression

M

mnarasim

What is the difference between Bi-Variate and Multi-Variate Regression.
What does LINEST do - a bi-variate or a Multi-variate regression.
How do we do both the kinds in EXCEL.

Thanks.
 
A

Alex

Bivariate means a regression using only two variables
(hence the word 'bi-' which means 'two') [I dont mean to
be patronising]

So multivariate regresion is more than two variables.

Usually Y is the dependent variable and X the independent
variable.

Bivariate example...

Suppose I think a persons height (X variable) is related
to their weight (Y variable). For example, if I think if
I know somebodys height then I can predict their weight
then I can do a bivariate regression of X on Y to find
out the strength of the relationship.

Multivariate Regression

Suppose now I think that knowing both height and age is a
better predicotr of weight. Here height and age are the X
variables and weight the Y variable. Here I have two
predictor (X variables) i.e. height and age. To find out
how they affect weight I need a multivariate regression.

For bivariate regression use simple Excel formulas such
as LINEST, SLOPE, INTERCEPT. An easy way is to plot your
data as a 'Scatter Plot' and then irght click the chart
and select <Trendline>. In the options tab select <Show
Equation On Chart> and this will give you your equation
for prdeicting Y from X in the form Y = mx + c

You cannot do the above for multivariate regression.
Instead you need the <Data Analysis> Add-In. (This is
free and may already be installed). Under <Tools> select
<Data Analysis...> and then <regression>.

By the way, I explained this below in your prior post.
Please recheck.

Also, doing regression is reasonably simple, but I would
advise you did a quick 'Google' and read abit beforehand.
There are pitfalls, and interpreting results is
notoriously difficult. Any standard statistic book will
have regression analysis. Snedecor 'Statistical Methods'
is a great starting point.

Regards


Alex
 
M

mnarasim

Yes, thanks. This makes sense. The reason I got confused is because the Help
for LINEST in EXCEL (Exaple 3) shows both 1) Bi-Variate and 2) Multi-Variate
regression. This made me think that LINEST could be used for both Bi-Variate
and Multi-Variate regression.

Alex said:
Bivariate means a regression using only two variables
(hence the word 'bi-' which means 'two') [I dont mean to
be patronising]

So multivariate regresion is more than two variables.

Usually Y is the dependent variable and X the independent
variable.

Bivariate example...

Suppose I think a persons height (X variable) is related
to their weight (Y variable). For example, if I think if
I know somebodys height then I can predict their weight
then I can do a bivariate regression of X on Y to find
out the strength of the relationship.

Multivariate Regression

Suppose now I think that knowing both height and age is a
better predicotr of weight. Here height and age are the X
variables and weight the Y variable. Here I have two
predictor (X variables) i.e. height and age. To find out
how they affect weight I need a multivariate regression.

For bivariate regression use simple Excel formulas such
as LINEST, SLOPE, INTERCEPT. An easy way is to plot your
data as a 'Scatter Plot' and then irght click the chart
and select <Trendline>. In the options tab select <Show
Equation On Chart> and this will give you your equation
for prdeicting Y from X in the form Y = mx + c

You cannot do the above for multivariate regression.
Instead you need the <Data Analysis> Add-In. (This is
free and may already be installed). Under <Tools> select
<Data Analysis...> and then <regression>.

By the way, I explained this below in your prior post.
Please recheck.

Also, doing regression is reasonably simple, but I would
advise you did a quick 'Google' and read abit beforehand.
There are pitfalls, and interpreting results is
notoriously difficult. Any standard statistic book will
have regression analysis. Snedecor 'Statistical Methods'
is a great starting point.

Regards


Alex
-----Original Message-----
What is the difference between Bi-Variate and Multi- Variate Regression.
What does LINEST do - a bi-variate or a Multi-variate regression.
How do we do both the kinds in EXCEL.

Thanks.


.
Click to expand...
Click to expand...
 
J

Jerry W. Lewis

More commonly, bivariate and multivariate refers to the number of y
variables to be simultaneously predicted.

You are correct that SLOPE and INTERCEPT only do simple linear
regression (one x-variable predicting one y-variable), but that is only
the beginnng of what LINEST can do. LINEST fits models involving a
single y-variable, but any number of x-variables that combine linearly
in the unkown coefficients (e.g. polynomials, etc.).

Jerry
Bivariate means a regression using only two variables
(hence the word 'bi-' which means 'two') [I dont mean to
be patronising]

So multivariate regresion is more than two variables.

Usually Y is the dependent variable and X the independent
variable.

Bivariate example...

Suppose I think a persons height (X variable) is related
to their weight (Y variable). For example, if I think if
I know somebodys height then I can predict their weight
then I can do a bivariate regression of X on Y to find
out the strength of the relationship.

Multivariate Regression

Suppose now I think that knowing both height and age is a
better predicotr of weight. Here height and age are the X
variables and weight the Y variable. Here I have two
predictor (X variables) i.e. height and age. To find out
how they affect weight I need a multivariate regression.

For bivariate regression use simple Excel formulas such
as LINEST, SLOPE, INTERCEPT. An easy way is to plot your
data as a 'Scatter Plot' and then irght click the chart
and select <Trendline>. In the options tab select <Show
Equation On Chart> and this will give you your equation
for prdeicting Y from X in the form Y = mx + c

You cannot do the above for multivariate regression.
Instead you need the <Data Analysis> Add-In. (This is
free and may already be installed). Under <Tools> select
<Data Analysis...> and then <regression>.

By the way, I explained this below in your prior post.
Please recheck.

Also, doing regression is reasonably simple, but I would
advise you did a quick 'Google' and read abit beforehand.
There are pitfalls, and interpreting results is
notoriously difficult. Any standard statistic book will
have regression analysis. Snedecor 'Statistical Methods'
is a great starting point.

Regards


Alex
-----Original Message-----
What is the difference between Bi-Variate and Multi-
Click to expand...
Variate Regression.
What does LINEST do - a bi-variate or a Multi-variate
regression.

How do we do both the kinds in EXCEL.

Thanks.
Click to expand...
Click to expand...
 
Top