Approximate Inference

Context

• Approximate Inference is a backup method for performing inference if creating a Junction Tree for Exact Inference is prohibitive in terms of memory requirement or computation time.
• Approximate Inference is based on the law of large numbers and uses network simulation to approximate the probabilities.
• While Approximate Inference initially requires very little memory compared to Exact Inference, each inference step with Approximate Inference takes time to perform.

Usage

• You can specify to use Approximate Inference by selecting Menu > Inference > Approximate Inference.