The Information Panel of a graph window is visible only in Validation Mode. It can be opened and closed with the button located at the bottom right corner of the panel. It contains up to six different statistics:
The Joint Probability corresponds to the current set of observations, i.e. hard positive/negative evidence, plus soft evidence. Without any evidence set, the joint probability must equal 100%. The joint probability is automatically updated with each modification of the monitors.
The Log-Likelihood is the logarithm of the Joint Probability to the base 2 . If the Joint Probability equals 100%, the Log-Likelihood is 0. Its tends towards infinity as the Joint Probability tends towards 0. Along with the Joint Probability, the Log-Likelihood is automatically updated with each modification of the monitors.
If a database is associated with the network, the information panel shows an estimate of the number of records that correspond to the current set of evidence, including hard positive/negative evidence as well as soft evidence. Without any evidence set, the number of cases is equal to the total number of records in the database. Whenever any evidence is modified via the Monitors, the number of cases is automatically recomputed.
If there is at least one node with associated values in the Bayesian network, the Information Panel displays the expected total value of all node with associated values in the network.
If there is at least one node with associated values in the Bayesian network, the Information Panel displays the expected mean value of all node with associated values in the network.
This value represents the uncertainty variations over all the unobserved nodes versus a fully unconnected network. This uncertainty value is computed from the entropy (for reference, a uniform distribution of states would yield the highest possible entropy for a node. If a state of a node is known with certainty, i.e. a 100% probability, the entropy is for that node would be zero). Please note that this uncertainty value is only computed if the corresponding option is checked in the Monitor Panel's Contextual Menu.
This value represents the Likelihood variations of the Bayesian network versus a fully unconnected network. The Likelihood is computed from the joint probabilities of the current set of evidence. The Likelihood is only computed only if the corresponding option in the monitor panel's contextual menu is checked
A tooltip displays all the previous values but without rounding them.