Hidden tiger (7): recursive Bayes formula (simulation, Matlab)

Antonio Sala, UPV

Difficulty: *** ,       Relevance: ,      Duration: 17:40

*Enlace a Spanish version

Summary:

This video presents a simulation of the recursive Bayes filter in the hidden tiger problem that we are analyzing in several previous videos. Specifically, the theory of the recursive filter was discussed in the video [tiger6brEN] (well, only adapted to this particular case study). Here, the code to carry it out is detailed and its results are discussed, obviously coinciding with those of the non-recursive formula in the video [tiger5EN].

Several simulations are made with different “signal-to-noise” ratios for the observations (percentage of roars heard from the wrong side). The noisier the observations, the more roars are needed to be sure where the tiger is actually hiding. In all simulations, the tiger always remains on the same side, as it is required to be so according to the assumptions we posed to carry out the statistical inference.

This is almost a full-fledged ”Bayes filter”, albeit the said full Bayes filter can incorporate the possibility of the tiger switching sides from time to time following a ”Markov process” dynamics with an associate transition probability table. These possible expansions are not considered here, for brevity.

*Link to my [ whole collection] of videos in English. Link to larger [ Colección completa] in Spanish.