Bayesian Filters Best Practices?

Configuration
1

I would like to configure two sensors, but I’m unsure on how to do it (if it is at all possible):

  1. Changing observations with time. For example, different probabilities exist in summer and in winter for “back door open when at home during the day” (in summer it is close to 1, in winter it is close to 0). Is it possible to integrate this into a single sensor “at home”, or is it necessary to create two sensors (“at home summer”, “at home winter”)?

  2. Look-ahead sensor/future event detector: in order to turn on the heating in the morning, I would like to create a sensor “likely to get up within 30 minutes” based on time of day, alarm clock, previous bedtime etc. Is there any generic or specific advice on how to create such a sensor, as it probably is mostly templating?

Any suggestions are appreciated. Thanks.