Hello,
I have a very basic question about a posteriori inference. Once evidence has propagated is it correct to take the weighted average of the mean values of the individual outcomes to get an exact value? Instead of the MAP value.
For example, if I had
s_10_20 (15)
s_20_30 (25)
s_30_40 (35)
Can I take the weighted mean by multiplying the a posteriori probabilities to the outcome values in parentheses. Would that more representative to get an exact value instead of just using the MAP outcome.
Getting exact Value from outcomes
Re: Getting exact Value from outcomes
You seem to be calculating the expected value. Depending on the task you're trying to perform this could be better than the MAP value.romeo wrote:Hello,
I have a very basic question about a posteriori inference. Once evidence has propagated is it correct to take the weighted average of the mean values of the individual outcomes to get an exact value? Instead of the MAP value.
For example, if I had
s_10_20 (15)
s_20_30 (25)
s_30_40 (35)
Can I take the weighted mean by multiplying the a posteriori probabilities to the outcome values in parentheses. Would that more representative to get an exact value instead of just using the MAP outcome.