What is the formula for calculating approximate costs using multiple variables?

[sergey]

Hello!
I have the following formula -
a*k1 + b*k2 + c*k3 = P
This is a formula I'm trying to build. It will be used to calculate price of a specific good (P) according to criterions k1,k2,k3. Every k have its own value, that's why its being multiplied by coefficients a,b and c.
I have a history of many previous purchases where k1,k2,k3 and P are known and i need to approximately calculate a,b and c according to this.

Hope i explained this well.
Can someone help me with this?

 

[atqamar]

Just an idea:

P is a function of k_1, k_2, and k_3.
So the multivariable formula is in the form P(k_1, k_2, k_3) = ak_1 + bk_2 + ck_3.

If you have many values of P and k in the formula P(k) = ak, you would run a single variable regression on the data to find a reasonable value of a. Thus far, I am not fluent in multivariable regression, so here is a http://www.nd.com/NSBook/NEURAL%20AND%20ADAPTIVE%20SYSTEMS17_Regression_for_Multiple_Varia.html" that might help.