In algebraic geometry an affine hypersurface is excactly what micromass said. There are no restrictions on the polynomial in this context, which means it doesn't need to be linear.
A geometric definition of a affine hypersurface in algebraic geometry could be "a closed subset of codimension 1 of an affine space". (an affine space is normally k^n, where k is algebraically closed field in the zariski-topology, or some subset of this if you want more generality) In other words it is a closed subset of an affine space with dimension one less than the affine space itself. Intuitively you can imagine a 2-dimensional surface (such as a plane, a sphere, a plane intersecting a sphere etc.. in euclidean 3-space). This means it is generated by a single polynomial. Sometimes a hypersurface refers to such an irreducible set (which is what I've seen, but I will not insist on this), which means that the generating polynomial needs to be irreducible.
Hypersurfaces generated by a linear polynomial are generally called hyperplanes (and specifically lines if the dimension is 1).