# Thread: type of regression model

1. ## type of regression model

Which type of regression model is it
y=a+bx+e

2. ## Re: type of regression model

In principle, it could be many things; it depends on the distribution assumed for the error term.

In practice, it is probably a linear regression (often mistakenly refered to as OLS; it is a mistake because Ordinary Least Squares is te algorithm used for finding the estimates, not the model itself).

3. ## Re: type of regression model

Can we say it normal equation

4. ## Re: type of regression model

Originally Posted by maartenbuis
In principle, it could be many things; it depends on the distribution assumed for the error term.

In practice, it is probably a linear regression (often mistakenly refered to as OLS; it is a mistake because Ordinary Least Squares is te algorithm used for finding the estimates, not the model itself).
thanks Maarten

5. ## Re: type of regression model

Originally Posted by sara b
Can we say it normal equation
I would not call it a normal equation. It is the simplest equation that can describes the relationship between two variables.

An equation can become steadily more complex by adding more IVs, then interactions between IVs, then quadratic terms, then cubic terms, and finally nonlinear terms.

6. ## Re: type of regression model

I have never heard the expression "normal" used for an equation. That describes a distribution of the residuals not an equation. Aside from the examples miner already gave equations can also have splines (although I guess those are non-linear).

7. ## Re: type of regression model

Originally Posted by noetsi
I have never heard the expression "normal" used for an equation. That describes a distribution of the residuals not an equation. Aside from the examples miner already gave equations can also have splines (although I guess those are non-linear).
thanks everyone for correcting my concepts

8. ## Re: type of regression model

Originally Posted by sara b
Can we say it normal equation
The following is usually referred to as the normal equation:

which is derived in the following way

.

A fancy thing with the normal equation is that you can show that .

.

9. ## Re: type of regression model

Originally Posted by maartenbuis
In principle, it could be many things; it depends on the distribution assumed for the error term.
Despite the distribution of the error terms we can say that the model is linear in its parameters, which is usually referred to when we're saying that a model is linear. The model

is also linear (in its parameters, that is). But the relationship between y and x is nonlinear.

10. ## Re: type of regression model

Exactly. This is always linear regression, irrespective of the values of X and distributional properties. The word "linear" stands for the dependence of y on b... Since there is only one predictor, many textbooks call this model "simple linear regression".

11. ## Re: type of regression model

Originally Posted by Englund
The following is usually referred to as the normal equation:

which is derived in the following way

.

A fancy thing with the normal equation is that you can show that .

.
Can you elaborate this further? I have an assignment on this where i need to prove that X'XB=X'y and B=(X'X)^-1*X'y for the given matrix as attached herewith.

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts