How Do I Fit the Equation x*y^a = b to Data in MatLab?

[2slowtogofast]

if i am given a data set of x and y values i need to come up with an expression of

x*y^a = b and b are constants

no idea how to get started any help geting started please

 

[viscousflow]

You may want to investigate the curve fitting toolbox

http://www.mathworks.com/products/curvefitting/description1.html

 

[2slowtogofast]

yeah i did polyfit(x,log(y),1)

that will return 2 number and ill have an eqn

lny = a1x+a2 it fit a straight line to my data but I am confused on what to do next i need xy^a = b is there another function i should try and use or is this more of an algebraic manipulation type of thing and if so can some one tell me where to start.

 

[viscousflow]

There is an algebraic way to solve that. Here is my approach

$$y^a = \frac{b}{x}$$

$$\log(y^a) = \log(\frac{b}{x})$$

$$a \log(y) = \log(b) - \log(x)$$

Since this is the homework forum I can't solve the entire problem for you. However, if you do some arranging you can do a $$y=mx+b$$ curve fit and solve for your coefficients.