Lecture 9

Question 1

Bayesian statistics meaning

Answer 1

1. Formulate competing hypothesis and assign a probability to them
2. Gather data about random variable of interest and evaluate to what degree observed data (dis)cofirm hypothesis
3. Update probabilities of hypotheses given the evidence

Question 2

Positive features, limitations and understanding between approaches to stats

Answer 2

Bayesian:
+ You can start with a prior belief
+ Flexible and adjustable

• Sometimes hard to pick a guess
• Sometimes hard to update and get firm answers

Like starting with a guess and upating as you go

Traditional (Frequentist):
+ No need for prior data
+ Same answer for everyone
+ Familiar and well-tested

• Less flexible
• Sometimes needs a lot of data
• Doesn’t tell confidence

you stick to the evidence and don’t assume anything extra.