VLIA: Navigating Shadows with Proximity for Highly Accurate Visited Location Inference Attack against Federated Recommendation Models