Back to practice exercises.

1: Background Reading

• 6.1 Probability
• 6.2 Independence
• 6.3 Belief Networks

2: Squeak Diagnostics

1. This Bayes net represents a bicycle squeak diagnostic for determining what might be causing a bicycle to squeak. Right-clicking on a node allows you to examine the probability tables for that node. Do the probabilities seem sensible? In plain English, describe the relationships between the nodes in this Bayes net.
   
   
2. What are the priors for MouseInSeat? What is the probability of the bike squeaking if MouseInSeat=T?

3: Making Observations and Queries

1. Let's say we made an observation that the bike is squeaking. We may have noticed this as we were pedalling home. Before querying the system, make a prediction about the most likely cause of the squeaking. Based on your answers in the previous section, do you think it's likely that there's a mouse in the seat given the bike squeaking?
   
   
2. Having written down your prediction, making an observation BikeSqueaks=T by navigating to the Solve tab in the applet, choosing 'Make Observation' from the panel and clicking on BikeSqueaks. Now choose 'Query' from the panel and query MouseInSeat. Are these values along the lines of what you expected? Why or why not?
   
   
3. Now make two more observations: ChainWorn=F and BrakesWorn=F. If you query MouseInSeat, has anything changed? Explain why or why not.