# Create a function from data points?

176 views (last 30 days)
A on 26 Jul 2017
Commented: A on 26 Jul 2017
Hi guys,
I want to create a function/equation from 10,000 data points. So I have three columns of input and one column of output. I want to be able to create a function y = mx+b type so that I can calculate the outcome with 3 inputs. This way, I want to use this y=mx+b function in something like excel to calculate an output given three input variables. Is this possible? How would I go about doing this? Here's an example:
Input 1: 5 Input 2: 15 Input 3: 0.5
Output: 20
So I want something like ... output = (Input1*2+Input2*2)*Input3
Thank you!
Image Analyst on 26 Jul 2017
output = (Input1*2+Input2*2)*Input3 is not y=mx+b with unknowns m and b that need to be computed, it's more like y=(2*x1 + 2*x2)*x3, except that you have no unknowns in that model. No coefficients to compute.
A on 26 Jul 2017
I understand. All I'm trying to do is .. I have 3 columns of numeric data... and 1 column of numeric output. I want to find a relationship between the 3 inputs and the output that gets resulted... and then I want to get a function that will help me get to that output using the 3 inputs. This way.. I want to put this function in an excel 'cell' and come up with output based on the data in the 3 columns? Thanks!

Santa Raghavan on 26 Jul 2017
Edited: Santa Raghavan on 26 Jul 2017
You can use the curve fitting toolbox to generate the model. Use this link to learn more: CurveFit
The toolbox provides the following models for mapping your input to output: Linear polynomial curve, Linear polynomial surface, Quadratic polynomial cu, Local linear regression (surface)
But however for multivariate (more than 3 input) linear fits, you can use the statistics toolbox. Here is a useful link: mvregress
You can also set up your own optimization problem to fit your function as: curve fitting via optimization