Maximum Likelihood Example: Normal

The use of maximum likelihood estimation to estimate the mean of a normally distributed random variable.

[@aidosmaulsharif9570]

Man this algebraic explanation of square deviations as quadratic equation was completely mind-blowing thing, instead of just taking derivative and equalizing it to zero, you simply solved quadratic equation. That was a great eureka moment for me. Thank you so much man!!! Subbed

[@BambooForestRecords]

I was so confused by my stupid notes. Thank you for explaining it so clearly. You are boss.

[@christopherpflaum8989]

I have been struggling to understand this since graduate school - now I do. Thank you.

[@TheRealRslive]

you sound like u just ran 100 miles

[@souslicer]

what about variance, how do i estimate that?

[@TheYeo1]

Great tutorial, very clearly explained. Thanks.

[@dchaitow]

Thats where RSQ comes from! Thanks for the video.

[@andrevauvelle4413]

An example including n samples with different variances would also be gr8!

[@muhammadhammadkarim8865]

to make things a bit simpler, where it is mentioned that minima of parabola is B/2*A, it would have been easier to say that the derivative is 0 at minima, so taking derivative w.r.t 'x' would yield the same thing. good video nonetheless.

[@Qbabxtra]

Intense breathing

[@schaeferconstantin]

Very, very good lecture!

[@f1mercury]

Excellent Demonstration. Thank you very much.

[@tangled55]

This is an EXCELLENT video on maximum likelihood estimation. Are you an actual professor?

[@GAWRRELL]

Can you please make an example of this using real world data?

[@study24816]

Thanks for your tut! but may I ask a question? you brought all together (1/sqrt(2pi*sigma))^n meaning all observations have the same sigma? how come?

[@TheYeo1]

@MathHolt Hi... I have some doubts. My question may be silly, but why we need to take log ? Can't we just differentiate and then assign to zero to get the maxima? Thanks.

[@MrArunavadatta]

well explained

[@gautama4590]

Great explanation... at the end we could have simply find derivative of the sum wrt mu and equate it to zero to find MLE of mu = x-bar

[@krish240574mumbai]

BEAUTIFUL

[@mrms154]

Breathing unbelievably

[@babyroo555]

genius!! thank you so much you have no idea how much you've helped me understanding this!! If I knew you i'd hug you :p

[@ukno918]

I've seen least squares with linear algebra using Ax = b and finding coefficients for the polynomial that'll regress through the data. Is this somehow related?

[@mkrasmus]

Great

[@satbirsaini8875]

It has to do with the monotonic nature of log, since log only changes the y values, the x value where the max occurs remains the same. Or at least, that's how my prof put it.

[@manmalhotra]

Good tut but if you had just taken the log and then differential to find maxima(of the whole -ve term), it'd have been a lot easier.

[@lilmoesk899]

One question: at 5:04 or so he says we can add the powers, but then writes a minus sign for all the powers of e. What am I missing?

[@study24816]

I see! so you concern about muy over sigma and assume all sigmas are constant and similar. Am I right?

[@MrArunavadatta]

nice

[@fakeplayer]

This heavy breathing........omg....

[@EvolutionCode]

AWESOME! :D

[@Rorkazak]

Did you just run a marathon ?

[@jmccoynpg]

\o/ MATH HOLT \o/

[@annalam8624]

thank you!!!

[@shrimatkapoor2200]

Why are you out of breath? Seems like a pretty strenuous calculation lol

[@JH-tc5wz]

why not just take dL/d mu = 0

[@pokirihoney]

i wish u were my teacher lol

[@mahesh-mh9hg]

marry me

[@visweshwaranra]

good video but his breathing is so much irritating