Perhaps a much more dense door is in order?
I am pretty sure bank vault doors don't blow in the wind, and are quite effective against burglarious attempts.
TO explain the bouncing in the direction where the door opens, assuming that the house has infinite mass:
1) the door is blown/sucked by the wind towards it's opening direction
2) the door hits the wall or the "end-stop" of the hinges.
3) the door still has considerable momentum
4) this momentum is transferred to the wall (or in the case of the hinge, stored as elastic potential energy)
5) due to a degree of elasticity between the wall and door (even though it isn't a rubber door!), or the energy stored in the hinge, momentum is transferred back to the door in the direction from whence it came.
6) because there isn't a 100% transfer of momentum, the door swings back with less force than it swung open
7) the wind catchs it again, blows it open.
8) because the wind applies a force to the door over a shorter distance now, the door absorbs less momentum from the wind than before
9) the door hits the end stops with less momentum than before, so bounces back even less
10) when the momentum required for the door to swing back against the wind becomes less than the momentum applied back to the door by the hinge or wall, the door remains open.
I hope that makes sense. The door blowing open probably has it's "bounce" in the internal air pressure of the house and the hinges not being perfectly level (as nothing ever is)