Linear Regression?

Adele69

Any suggestions on a good online or book resource which covers this?

My Osbourne Financial Performance book skimmed over this - equation y = a + bx, a line on a graph, Line of Best Fit, Least Squares and mathematic formula not covered, and Budgeting just repeated the earlier content. I can calculate some of these but also get stuck on others. Usually the wording catches me out, so suggesting I don't really understand what is being asked.

Bertie

Which is it you need help with? All mentioned?

Adele69

I can work out values of a and b for questions which give two examples for y and x using high-low, and other time series where a period is all but given. However a scenario with several months where high-low proves it isn't a regular variation is...? so my guess is either a line of best fit (never done a question on this) or least square (dont think I've done one of these either). There are no worked examples (Case Studies) for these so really don't know how to identify them.

Bertie

Ok, I'm not sure how this will look in a post -

Least squares regression -

To forecast sales make the dependant variable Y and the number of sales people the independent variable X.

So first, some data -

Sales units - Salespersons
236. 11
234. 12
298. 18
250. 15
246. 13
202. 10

So now you need to calculate the sum of, (I can't type the sum of character of sigma, so I'll use E to mean such, also 2 means power of)


So,
Ex 79
Ey 1466
Ex21083
Exy 19736

To calculate B,

n = number of data sets, in this case, 6.

B=nExy - (Ex)(Ey) / nEx2 - (Ex)2 = 10.12

Now must find the average of X and Y

So, X = 13.17
Y = 244.33

A is therefore = mean of Y - B(Average of X)

A = 244.33 - 10.12(13.17) = 111.05

So the least squares of regression is

Y = 111.05 + 10.12x