[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math

View post   
File: 19 KB, 354x400, gm_regression.jpg [View same] [iqdb] [saucenao] [google]
5800507 No.5800507 [Reply] [Original]

What is the best book to learn regression analysis and other statistics related subjects? Also how does one getter better at analyzing Stata data?

>> No.5800515

The Manga Guide to Statistics

>> No.5800524

>>5800507
>x vs y
>not a log log plot
>hand drawing a line

gtfo

>> No.5800525

>>5800524
You jelly?
https://www.youtube.com/watch?v=l789l6np-qA

>> No.5800577
File: 1.86 MB, 3264x2448, 20130531_142312.jpg [View same] [iqdb] [saucenao] [google]
5800577

Pic related. Applied Linear Statistical Models if you want even more.

>> No.5800607

>>5800507
OCR S1

>> No.5800736

I'm less curious about linear regression and more curious about best fit functions in general, not just linear.

Mathematica, for example, can evalutate Fit[Data, {1, x, x^2}, x], which will give a parabolic approximation, rather than a simple linear one.

>> No.5800747

>>5800507
If regression is finding a curve to approximate a set of values, then it doesn't sound that hard. Just classify the sort of curve that you want to get, for example ax+b for a straight line or ae^(bx) for an exponential curve, then minimize some kind of function that you define to represent how far away from the optimum your curve is, for example the average distance between the data points and the curve, to find the optimal paramaters (a, b, etc.)

>> No.5800749

>>5800525
I FUCKING LOVE CHALK

>> No.5800756 [DELETED] 

>>5800507
>x on the vertical axis
>y on the horizontal axis