Tags :: Bayesian Statistics

Have two defining characteristics:

• Unknown quantities are described using probability distributions (parameters)
• Bayes theorem is used to update the values of the parameters conditioned on teh data. Also a reallocation of probabilities.

Process of constructing modeling

1. First we have assumptions on how data has been/could have been generated. For this we design a model by combining and transforming random variables.
2. We use Bayes theorem to condition our models to the available data. This is called inference (Bayesian Models). From this we receive the Posterior distribution. We hope the data reduces uncertainty for possible parameter values, though this is not a guarantee of any Bayesian model.
3. Criticizing the model against our data and expert domain knowledge allows us to determine whether or not the model makes sense. Typically you will want to compare multiple models as we should generally be uncertain about the models themselves