I dont know either, so just another speculation:
If you imagine a ball on the end of a string, if you knock it, it will oscillate. If you were to knock it again in the middle of its oscillation, it would just oscillate again based on the 2nd time you knocked it (with the same frequency, but different phase effectively)
if you were to knock it again exactly as it finished its first oscillation however, it would swing further, or oscillate with a greater amplitude. Thats effectively forcing it at its natural frequency - causing resonance.
It'll always oscillate at a certain frequency (its natural frequency), no matter how hard you knock it, as its based on the length of the string (and other factors?) which i suppose basically effect the forces acting on the ball.
So generally you can't really model it mathematically as you cannot determine all of the forces holding the object in place, or it'd just get ridiculously complicated?
just speculation again I'm afraid though - its an interesting question