Imagine that you have a computer with the following properties:
flat
black
hot
speck of red on one side
And your prior at this point for p(slow given it’s a computer) = 0.5.
If an article about computers said p(slow given flat, black, and speck of red) = 0.25, then would you use the new # as your prior or combine the two pieces of information to calculate p(slow given flat, black, hot, and speck of red)?
I’m inclined to say that I should use .25 as the new prior and forget about 0.5 but I may also be making a silly logical error here.
I wouldn’t call this trivial, it depends on the particulars (which is why in the notation above, there’s a contradiction around the value of p_1). The easiest resolution involves following up on the article by looking for data/causal models. Like using the computer, and acquiring evidence.
Potentially trivial math question:
Imagine that you have a computer with the following properties:
flat
black
hot
speck of red on one side
And your prior at this point for p(slow given it’s a computer) = 0.5.
If an article about computers said p(slow given flat, black, and speck of red) = 0.25, then would you use the new # as your prior or combine the two pieces of information to calculate p(slow given flat, black, hot, and speck of red)?
I’m inclined to say that I should use .25 as the new prior and forget about 0.5 but I may also be making a silly logical error here.
o_1 = (observed: flat, black, with red speck)
p_1 = p(slow given o_1) = 0.5
o_2 = (observed: article claiming p_1 = 0.25)
p_2 = p(slow given o_1, o_2) = ?
I wouldn’t call this trivial, it depends on the particulars (which is why in the notation above, there’s a contradiction around the value of p_1). The easiest resolution involves following up on the article by looking for data/causal models. Like using the computer, and acquiring evidence.