Your polyfun requires two inputs. Your pf takes a handle to that directly, so pf will be a handle to a function that expects two inputs. You then call fplot() which is documented as
The function must be of the form y = f(x), where x is a vector whose range specifies the limits, and y is a vector the same size as x and contains the function's value at the points in x (see the first example). If the function returns more than one value for a given x, then y is a matrix whose columns contain each component of f(x) (see the second example).
We can see that the function is expected to take only one input, so fplot() is only going to be trying to pass one input. But your function needs two inputs so it is going to fail.
What you need is to have defined your polynomial, such as
x = sort(rand(1,10)); %some arbitrary data
y = sort(rand(1,10));
p = polyfit(x, y, 4); %create polynomial coefficients
and then ensure that this p is passed into polyfun:
pf = @(x) polyfun(x, p);
and then you can
fplot(pf, [-0.5 0.5])
Notice that the function handle being passed to pf only expects one input, and it takes that one input and expands it into a call to polyfun providing a pre-set value p for the second input to polyfun . fplot is happy because the function it called only needed one argument, and polyfun is happy because it received two arguments.
