This work develops a method for learning about the meetings in others calendars based on their responses during the process of meeting scheduling. We gradually build up a probability distribution of meeting types for each possible timeslot in another users calendar. As we gain more information about other users calendars through additional negotiations, for example over the course of a semester, this distribution converges on the other user's actual calendar. This information can be used to speed the meeting negotiation process or to help users schedule meetings to their own benefit. This work is designed to integrate into the CMRadar project.