Bayesian statistics
5 important questions on Bayesian statistics
What is necessary to perform bayesian statistics?
- Data
- a generative model
- priors; what information the model has before seeing the data
What is bayesian inference?
- A generative model can be used to feed parameter values and generate data.
- often we are in the opposite situation; the data is known
- in bayesian inference you want to use the generative model to see what parameter values can give rise to the found data.
What are ways to summarize the posterior probability distribution?
- the maximum likelihood estimate is the parameter value that is most likely to result in our found data.
- the posterior mean is the mean parameter values that is most likely to result in our found data.
- a x% credible interval; the chance of the parameter being between these two boundaries is x%
- Higher grades + faster learning
- Never study anything twice
- 100% sure, 100% understanding
How can you use the bayes factor to interpret data?
- A bayes factor of 1 indicates that the experiment did not provide any evidence towards h0 or h1
- a bayes factor >1 indicates that the data provides evidence for h1
- a bayes factor <1 indicates that the data provides evidence for h0
What are advantages for bayesian statistics compared to frequentist statistics?
- P values do not give evidence for H0 regardless of whether power is high or low. The bayes factor does have the ability to give evidence for H0
- P values do not provide evidence for H1 in ways sensitive to properties of H1: P values do not punish theories for being vague, bayesian statistics do this.
- bayes factorsprovide a continuous measure of evidence based on the data, knowledge can continuously get updated.
The question on the page originate from the summary of the following study material:
- A unique study and practice tool
- Never study anything twice again
- Get the grades you hope for
- 100% sure, 100% understanding