Exercise 3.18 - Bayes factor for coin tossing

Answers

The Bayes factor for hypothesis test is defined by:

BF 1 , 0 = p ( data | H 1 ) p ( data | H 0 ) ,

where H 0 is the null hypothesis.

We have:

p ( data | H 0 ) = Bin ( 9 | 0.5 , 10 ) = ( 10 9 ) 0 . 5 10 0.00977 .

And

p ( data | H 1 ) = 1 10 + 1 0.09091 ,

according to Exercise 3.17. The Bayes factor is approximately 9.3 .

When N = 100 and N 1 = 90 , the Bayes factor is:

1 100 + 1 ( 100 90 ) 0 . 5 100 > 2 100 10 1 11 113622530 .

When N N 1 remains a constant deviated from 0.5, the larger N is, the more likely that the coin is biased. This is an intuitive conclusion from the law of large numbers.

User profile picture
2021-03-24 13:42
Comments