, Duration: 20:34*Enlace a Spanish version
Materials: [ Cód.: Tiger1ENG.mlx ] [ PDF ]
This video is the second one of the “hidden tiger” case study, which began in the
video [
In the aforementioned video the problem was posed, and a probabilistic model was defined. It was based on a probability of a left/right tiger after closing a certain gate and a conditional probability of which side (left/right) it is heard roaring based on (conditioned to) which side the tiger remained trapped in. With this, the joint probability table was obtained.
In this video, the following probability tables are obtained from the joint probability one:
Marginal of ‘tiger location’ (recovers the starting prior probability)
Marginal probabilities of ‘side where roar is heard’
Conditional from tiger location (condition) to roar hearing side (conditioned variable); We already knew this one, we are just tracing back what we did earlier on.
Conditional from roar listening side (condition) to tiger location (conditioned variable)
The first marginal and conditional have the meaning discussed in the first video; The second marginal tells the “accumulated frequency” of left or right roars if the experiment were repeated many times.
The second conditional table has a different meaning; roughly, the arrow of
causality goes “backwards”: given an observed effect (roar), the posterior
probability of its two possible causes (tiger on the left or tiger on the right) is
inferred. This idea is the basis of Bayesian statistical inference. The following
video, [
*Link to my [ whole collection] of videos in English. Link to larger [ Colección completa] in Spanish.