Restaurant Bayes

The story of Restaurant Bayes is as follows: One evening you go to Restaurant Bayes. Unfortunately, the food is of low quality and you would, therefore, like to reason which chef prepared the food of low quality. Based on this experience you would like to infer who prepared the food that evening.

The story continues: Peter and Eric are the chefs at Restaurant Bayes. Peter works six days a week while Eric works one day a week. In 90 per cent of the cases Peter's food is high quality while Eric's food is high quality in 50 per cent of the cases.

Given our knowledge of the problem is it fair to conclude that Eric prepared the food of low quality? We can use a Bayesian network model to represent the knowledge and support the reasoning process.

The Bayesian network above specifies the relation between the chef and the quality of the food. The strength of the relation is specified based on the knowledge we have from above, i.e., the prior probability of Chef being Peter is six out of seven and he prepares food of good quality in ninety per cent of the cases.

Demonstration of the model

Select Quality of Food

Possible Chefs Probability Probability (bar view)

Interact with the form to compute the probability of Eric being the chef once we observe the quality of the food to be awful.