Greater precision with regression?

I utilized my TI - 83 to locate the square regression of 2 information columns. The precision had not been close in all. So I attempted cubic and afterwards ultimately quartic regression. The precision still isn't close sufficient. Exists any kind of means I can boost the precision? I can not add anymore information, unfortunetly.

Modify 1:

To resolve Henry is comment, my outlined information looks comparable to y = x ^ (1/3)

2022-06-07 14:34:15
Source Share
Answers: 2

This might suggest your information is inadequately fit by a polynomial. 3 means it can fall short is either to be rapid, to have posts, or resemble a sine wave. Outlining the information with the fits you have actually superimposed might offer you some suggestions. If it is rapid, taking the logarithm of one will certainly provide it straight (or near polynomial). If it has posts, making use of a sensible function will certainly aid. Treatments are given up phase 3.2 of Numerical Recipes. The out-of-date variations are free. If it is a sine wave, an FFT can aid - phase 12 of NR.

2022-06-07 14:56:05

If, as you claim, it resembles $y=x^\frac{1}{3}$, after that you could attempt a power regression, which will certainly offer you a formula of the kind $y=ax^b$.

2022-06-07 14:53:31