--- In Statisticians_group@..., Sonia Singla <sonia203@...> wrote:
>
> Hi
>
> I am fresher in Regression Analysis and i have not any practical knowledge
> about this.
> Suppose we have a data and we want to fit a linear regression model then
> what are the points which we should keep in mind and what is the process of
> fitting a model.
>
> Sonia Singhal
>
>suppose you have data on two variables x and y
>first you calculate correlation between two variables or just plot the
scatterplot. if they are linearly related then we can fit linear regression as
follows
>here we assume model y=a+bx+e. a and b are parameters ,e is an error term. we
estimate these parameters using ordinary least square method(ols) as follow
>in ols method ,we minimise error sum of square,i.e sum of (y-a-bx)^2
over all n points
s=sum of (y-a-bx)^2 over all n points
>differentiate this s w.r.t. a and b then you get two normal equations and
solving these two equations,you get ols estimate of a and b