Harvard statistics, video 4 note(conditioning & conditional probability)

9:01 2014-10-04 Saturday
start Harvard statistics, video 4


conditioning & conditional probability


9:01 2014-10-04
independence:


Definition: Events A, B are independent if P(AB) == P(A)P(B)


9:20 2014-10-04
independent, disjoint


// completely different from disjoint


9:21 2014-10-04
disjoint: if we know A occured, then B cannot possibly occur


independent: if we know A occured, it tells us nothing whatsoever 
   
             whether B occured or not


9:23 2014-10-04
independence, pairwise independence


9:26 2014-10-04
"indepence means multiply when comes the probability of intersection"


9:27 2014-10-04
binomial probability


9:46 2014-10-04
conditional probability


9:53 2014-10-04
conditional probability:


How should you update your probability/beliefs/uncertainty


based on new evidence.


9:57 2014-10-04
"conditioning is the soul of statistics"


9:58 2014-10-04
P(A|B): probability of A occurs given B occurs


9:59 2014-10-04
we no longer assume naively that each outcome


is equally likely.


10:03 2014-10-04
event is a subset


10:05 2014-10-04
P(A|B) all we have to do is get rid of all pebbles


that are not in B  // in this smaller universe


10:06 2014-10-04
this is the correct normalization


10:09 2014-10-04
* pebble world


* frequentist world


10:12 2014-10-04
can you do this experiment over & over again?


this is a deep philosophical question.


10:13 2014-10-04
we just keep going like that


10:20 2014-10-04
I would like to relate P(A|B) & P(B|A)