- #1

- 159

- 0

## Main Question or Discussion Point

Like the title says, if you have a bunch of data you can create a best fit line

For example back in school, for a linear line, y = mx + c, you just need 2 points to get a line. But what if you have more than 2 points, say 10, what is the best fit line equation or how do you do it mathematically.

I'm sure excel can do it through the trendline method but back to basics, how do people do it. I'm looking more for polynomial, 2nd order. And data is not like 1,2,4,8,16 which you can deduce to y = x^2. But more like 'double' or 'float', so trying to get a best fit line and future prediction is a lot more harder

Just wondering if anyone ever look at something like this or how people find coefficients based on data information they have accumulated (which is how in real life happens and then deduce your own equation to reflect the change in information)

Is there a book somewhere or resources i could look at would be helpful too

Thanks

For example back in school, for a linear line, y = mx + c, you just need 2 points to get a line. But what if you have more than 2 points, say 10, what is the best fit line equation or how do you do it mathematically.

I'm sure excel can do it through the trendline method but back to basics, how do people do it. I'm looking more for polynomial, 2nd order. And data is not like 1,2,4,8,16 which you can deduce to y = x^2. But more like 'double' or 'float', so trying to get a best fit line and future prediction is a lot more harder

Just wondering if anyone ever look at something like this or how people find coefficients based on data information they have accumulated (which is how in real life happens and then deduce your own equation to reflect the change in information)

Is there a book somewhere or resources i could look at would be helpful too

Thanks