/sci/entists I need your help...
How the fuck did this guy go from (2.40) to (2.41). Please spell it out for me I am retarded.
>>8985389
might be a typo
[eqn] \prod_{j=1}^{|x_i|} q_{x_{ij}}(x_{ij}) \ln \prod_{j=1}^{|x_i|} q_{x_{ij}}(x_{ij}) \\
= \prod_{j=1}^{|x_i|} q_{x_{ij}}(x_{ij})\sum_{j=1}^{|x_i|} \ln q_{x_{ij}}(x_{ij})\\
=q_{x_{i}}(x_{i})\sum_{j=1}^{|x_i|} \ln q_{x_{ij}}(x_{ij}) \\
= \sum_{j=1}^{|x_i|}q_{x_{i}}(x_{i})\ln q_{x_{ij}}(x_{ij})[/eqn]
>>8985405
Yes I also reached the same last line. Thank you based anon. I would hope that there isn't a type like this in a dissertation..
Please any other summa brain polyglots weigh in.
Idiots are acting like they know more than me and calling me dumb but I'm at home while they're at summer school
>>8985409
Thank you for the bump based shitposting-bot.
>>8985408
There's typos everywhere
>>8985414
True and sad
>>8985389
Surely there are some Bayesians on /sci/ right...?
>>8985389
come on you guys always call each other brainlets... surely at least one of you is NOT a brainlet right?
>>8985389
I think it has to be a typo. The last product is killed by the properties of the log. This leaves the expression of the sum in the second line. This ensures that there is nothing to kill the product in the first line, so it has to be a typo.
>>8985389
Mad props anon for studying mft. Way of the future right there.