How MLE (Maximum Likelihood Estimation) algorithm works

[@bharathshashankakatkam255]

After watching, reading roughly about 150 videos & articles, now I finally understand the concept with the help of your lovely explanation. Thanking you desperatelyyyyyyy...

[@tkorting]

+Bharath Shashanka Katkam thanks for this good feedback please like and share the video and subscribe to my channel regards

[@sahildadia590]

Machine Learning Just became so much clearer after seeing this video. Best explanation ever!!!!

[@tkorting]

thanks for your feedback Please subscribe to my channel and share this video with your peers Regards

[@bombusaffinis]

Thank you very much for this explanation. I only use MLE as an end user, and don't really need detailed knowledge of every little bit and piece, so this was perfect for me. Thank you!

[@tkorting]

thanks for your feedback. please like the video, subscribe to my channel and share with your peers. Regards

[@kingpopaul]

Great visuals and nice in depth explanations. I find that technical things get overly complicated with specialized vocabulary that represent simple things that are easily forgettable, so these video really helps to remember and learn. Hats off to you good sir, I'll stick around for more!

[@tkorting]

+kingpopaul thanks for your feedback. This is my purpose, help people to remember the tricks of the algorithms using images and animation. Please subscribe to my channel and like/share the video. Regards

[@TheRealDCoy]

This is the most straight-forward basic introduction I have seen. Thanks and great job!

[@tkorting]

Thanks a lot for your feedback! Please subscribe to my channel and like the video. Regards

[@vuminhquanle1426]

Thanks man, solid explaination. Knowledge of normal distribution function and basic probability should be enough to understand this

[@tkorting]

Thanks for the comment, please subscribe to my channel

[@kiranvadlamani5807]

This is the 1st video I ever saw about MLE. Lucky me! Amazing!

[@tkorting]

Thanks for the feedback, please subscribe to my channel

[@vitid1]

The most lucid explanation about MLE

[@tkorting]

+vitid1 thanks for your feedback, please subscribe to my channel and share this video with your peers. Regards

[@ibal6875]

Excellent explanation regarding classification problems. Thank you sir!

[@tkorting]

Thanks for your feedback. Please like and share the video and subscribe to my channel. Regards

[@Naru1243]

Thank you so immensely,

[@tkorting]

Thanks for the feedback, please subscribe to my channel

[@fortoday04]

Wow. I am blown away by the quality of this explanation. Although your presenting voice could use a tiny bit of improvement - really, this is the only thing I can pick on - the entire video was superb. Well done!

[@aravindsairam3432]

This video is seriously underrated.

[@tkorting]

Thanks for the comment, please subscribe to my channel

[@justpaperskw]

Thanks, it is a really great and straightforward explanation for MLE. Liked.

[@tkorting]

+justpaperskw thanks for your feedback. Please subscribe to my channel. Regards

[@donghun740504]

Nice and clear explanation for MLE...

[@tkorting]

+김동훈 thanks for the positive feedback. Please like and share the video and subscribe to my channel. Regards

[@kevinsong1056]

Best explanation about MLE ever!!!!!!

[@tkorting]

thanks for your feedback, please subscribe to my channel and share this video with your peers regards

[@tarunjain07]

Awesome explanation. Very intuitive. Thank you!

[@tkorting]

+TARUN JAIN thanks for the positive feedback Please like and share the video and subscribe to my channel. Regards

[@christopherpestana212]

Brilliant explanation. Finally understood this.

[@tkorting]

Thanks for the feedback, please subscribe to my channel

[@SergioSotoQuintero]

Amazingly simple! Thank you so much!

[@tkorting]

Thanks for the feedback, please subscribe to my channel

[@viccctv9106]

Thanks for this so well explained video. Cheers from Bolivia

[@tkorting]

Cheers from Brazil Please like/share/subscribe Regards

[@redherring0077]

Wow! Such an intuitive video. You are an excellent teacher. Please keep it up! Liked and subscribed💜

[@Marco-wz8xw]

Great video and explanation. Made the subject super simple! Thanks.

[@tkorting]

Thanks for your positive feedback! Please like/share and subscribe to my channel Um abraço :)

[@valeriipotokov9094]

Great! It's very interesting subject and the presentation is very clear and practical. Thank you also for book reference at the end of this video.

[@tkorting]

Dear Valerii Potokov thanks for your valluable comments. and for subscribing to my channel. Regards

[@pascal3934]

Thank you Thales. One of the greatest explanation made on youtube. Keep it up please

[@tkorting]

thanks for your feedback, please subscribe to my channel and share this video with your peers. Regards

[@ganeshcsaiisc]

Thanks, Thales! That was an easier explanation!

[@tkorting]

Thanks for your feedback Please subscribe to my channel Regards

[@husamalfergani7110]

Thank you. Simple and easy

[@tkorting]

Thanks for your feedback Please subscribe to my channel Regards

[@Dba242]

Best explanation i've seen on yt

[@tkorting]

thanks a lot for your feedback please subscribe to my channel and share this video with your peers regards

[@조동민-f6o]

Great works...! I grasp the concept! Thanks a lot :)

[@tkorting]

Thanks for your feedback please like and share the video and subscribe to my channel Regards

[@tusharkantirouth5605]

Simple and to the point.... love this one after watching a ton of bullshits...

[@tkorting]

Thanks for your feedback Please subscribe to my channel Regards

[@muhammadumairkhan8435]

Great explanation. Thank you very much. I was stuck exactly at this point..... Hats off!!!!!!!!!!!!!!!!!!!!!

[@tkorting]

Thanks for your positive feedback, please like and share the video and subscribe to my channel. Regards

[@murt4701]

Great visuals, you boss

[@tkorting]

thanks for your feedback please like/share/subscribe Regards

[@gauravms6681]

just an amazing video

[@AsadKhan-bi5jg]

Very well explain with example Hats Off from Pakistan :)

[@tkorting]

Thanks for your feedback. Please like the video and subscribe. Regards

[@Leon-pn6rb]

Please don't learn this to make bombs thanks from Australia :)

[@hitendramotwani4793]

:D :D :D

[@samuelfsmith2134]

you explained this so well i had to sign in and comment.

[@tkorting]

thanks for your feedback, and for subscribing. Regards

[@jbberinger]

Great explanation!

[@tkorting]

Many thanks for your positive feedback Please like/share/subscribe. Regards

[@bharathshashankakatkam255]

Awesome Sir!!!

[@tkorting]

Thanks for the comment, please subscribe to my channel

[@fernandoarancetagarza8164]

Hello. In the last part of your video I notice in your formula p(x|w_j) for the blue class data you put a sigma for x and y different from your color class descriptive estimation in the minute 4:17. Why is this difference? your MLE formula says is a sigmax = 1.04 and sigmay = 1.03, even though in your color class descriptives you put 1.08 and 0.90. thank you

[@tkorting]

Hi Fernando, thanks for your feedback. You are right, this discussion is already present because due to running 2x the algorithm with random initial parameters, unfortunately when I created the video I did not perceive this small difference. I will try to insert a comment in this part. Regards

[@treeforty6734]

Thales muito boa sua explicação! Obrigado!

[@tkorting]

Muito obrigado pelo retorno! Inscreva-se no canal ;)

[@o.alejandrobauernfeind2062]

Very well explained

[@tkorting]

Thanks a lot for your feedback, please subscribe to my channel and share this video with your peers. Regards

[@feraudyh]

Gosh, it was *that* simple!

[@tkorting]

thanks for your feedback, please like/share the video and subscribe to my channel regards

[@scargomez9437]

Actually I am taking the Mchine Learning Course in edX of Columbia University, I am utterly lost in this subject. Thanks for the explanation, do you have a recommended list of books for beginners?

[@tkorting]

thanks a lot for this comment! please like/share the video and subscribe. Regarding the references, in the end of my videos, I usually put the references I used. Most of them are very good, and complete. However, the beginning is always harder, but then you get the concepts more easily. Regards

[@kalebdr1]

Nice video. Could you please upload video for probabilistic PCA ?

[@tkorting]

thanks for your feedback and suggestions. I will put it on my list for future videos. please like/share the video and subscribe regards

[@arihantparsoya]

Very good explanation

[@tkorting]

+Arihant Parsoya thanks for the positive feedback. Please like and share the video and subscribe to my channel. Regards

[@ЄвгенійГизила]

Thank you. Good explanation!

[@tkorting]

Thanks for the good feedback. Please like/share the video and subscribe to my channel. Regards

[@chrishanni2779]

Awesome, thank you

[@tkorting]

Thanks for the feedback. Please like and share the video and subscribe to my channel. Regards

[@maheshdusane6115]

Thank you, you really have made it very simple. It was a great help. I would like to get more in-depth of the class conditional density function (introduced @4:35), please give me some pointers where it explained in easy way in further details.

[@petraschubert8220]

How you got the 1,04 and the 1,03 standard deviation? And how you found the mü`s -3,89 and - 2,89? Thanks a lot for your help!

[@tkorting]

Thanks for your feedback, please like and share the video and subscribe to my channel. Take a look at previous messages in which I explain that I created a random set of values, and for some reason the first distribution is a little bit different from the second, and this can cause some doubt. But to get the values, I simply take all points and compute the average value for both x and y axes. Regards

[@liyingduan9948]

Thanks for explanation.

[@tkorting]

thanks a lot for your feedback Please subscribe to my channel and share/like this video! Regards

[@sathya_official3843]

awesome work! just misses a bit of background info though.

[@ajaydhungana1921]

a simple question..sir..how does maximum likelihood estimate is Related to variance components..I do know MLE estimates best parameters (mean and variance) from data...but where does variance components come into play. What is the link??

[@kimkieu3010]

When you calculated p(Xx,omega_blue), where is 1.04 and -3.98 come from? I thought mux_blue = -3.88 and sigma_blue=1.08 as given? Thank for the video. It's very helpful!

[@tkorting]

+Kieu Nguyen dear friend, thanks for your valuable comments. When I created the octave routine to make the animation at 5:01, I set random values to be around the tree clusters. And when I wrote the calculations of probabilites, I started the algorithm again, so there is a slightly difference between both graphics (see video at 3:51 and 5:01). So you are right, considering the distributions showed at 3:51, mux_blue should be -3.88 and sigma_blue should be 1.08. Thanks for your feedback. Please subscribe to my channel and share this video with your peers. Regards

[@kimkieu3010]

+Thales Sehn Körting By the time you replied my comment, I already finished my Pattern Recognition class xD Thank you anyway!

[@shashidharrapolu]

Kieu Nguyen h

[@AslamShaikh-oi9vr]

@@tkorting Thanks I had the same question. Answered.

[@iidtxbc]

Thank you for the awesome example. I just wonder if you take the independence assumption between x1 and x2 variable for computing probabilities.

[@Logaff]

really good video :)

[@tkorting]

+Logaff thanks for your feedback please subscribe to my channel and share this video with your peers. regards

[@Global_Info]

Very nice. I appreciate

[@tkorting]

Thanks for the feedback, please subscribe to my channel

[@rzipper1716]

Thanks ! Well Done

[@tkorting]

Thanks for your feedback. Please like the video and subscribe ;) Regards

[@TheKaraqi4]

what is the relation between densities and z values when calculating ML? how do we choose densities e.g. with Null Model? Can I find lessons about it somewhere?

[@ei256]

I salute to you!

[@tkorting]

+E I I also salute to you. Please subscribe to my channel. Regards

[@studentcommenter5858]

Seriously is that it? My book seems to have a lot of gibberish maths which was very difficult to follow. Thanks a lot.

[@musiichaven7349]

the best explain one question please why should take for y not only for x only ? thanks

[@tkorting]

+Musiic Haven thanks for your feedback, please subscribe to my channel and share this video with your peers. In this case we are using a 2-dimensional example, therefore we need to compute for both x and y. Regards

[@Husamadin]

Thanks for great explination. My question is, what if we have feature space with 3D like X, Y and Z??

[@tkorting]

Thanks for your question. Please like and share the video and subscribe to my channel. In case of more variables (I used only 2 because it is easier to display on screen) you will add z values in the equations, exactly the same way as with x, and y. Regards

[@kushalneo]

good work (y)

[@tkorting]

+kushal kanwar thanks for your feedback, please subscribe to my channel and share this video with your peers. Regards

[@kushalneo]

Nitish Mahajan watch it.

[@kushalneo]

Thales Sehn Körting sure

[@yurchax]

At 2:05 you say that's how we calculate the standard deviation, but isn't that the Covarience formula? Am I missing something?

[@wyphonema4024]

6:05, how did you get 1.04 and 3.98?

[@tkorting]

Thanks for the feedback, please subscribe to my channel Please check the previous comments, where I explain the inconsistency from the previous slide (I used random numbers for each slide)

[@samrabiba6867]

thnik's Mr Tales it 's verry nice this vidéo you are help me -chebeb Med -

[@tkorting]

thanks for your comment please subscribe to my channel and share this video with your peers regards

[@siddharthrawat7205]

What if the features have higher dimensionality? how to implement MLE in that scenario?

[@usmansaeed678]

Salute to you

[@tkorting]

Thanks for your message. Please like and share the video and subscribe to my channel. Best regards

[@mateushabermann4400]

Muito bom!

[@tkorting]

+Mateus Habermann muito obrigado pelo comentário. Por favor inscreva-se no canal, de like e share com seus contatos também. Um abraço

[@Akash-kq8lm]

is this also called gaussian mixture model?

[@ahmettarktorun4174]

Hi,it is so helpful video; can u help me about matlab maximimum likelihodd hyperspectral image classification

[@tkorting]

Hi Ahmet, thanks for your comments. Please subscribe to my channel and share this video with your peers. Regarding matlab, I would suggest you to google for it, certainly several tricks can be found. The basic idea is to open the image as a matrix, which will have 3 dimensions, 1 is the number of rows, 2 is the number of columns, and the 3rd dimension will be the bands of the hyperspectral image. Regards

[@doanthivanbinh2]

Hi Thales Sehn Korting Is C or (ω_i) the number of classes? Thank you!

[@tkorting]

Hi, thanks for your question Please like/share the video and subscribe. C is the number of classes. Then w_i is the description of each class, considering i = 1, 2, ... up to C Regards

[@dondan2504]

So what if we have 3 dimension? How to calculate the pdf?

[@tkorting]

Hello, thanks for your feedback. Please like/share and subscribe. The idea is the same. For example, around 6:35 you can see the way to compute the probability for x and y. If you have z, you just compute this as well, and the final probability will be px * py * pz, and so on. Regards

[@dondan2504]

Thales Sehn Körting Thank you so much for your reply! I liked and subscribed! I'm so addicted to machine learning. I think it became my religion lol!!!

[@tkorting]

great! thanks for subscribing

[@jianxiongji2578]

thank you

[@tkorting]

thanks for your feedback please subscribe to my channel, like/share the video regards

[@cokaneytp7168]

Very useful video, but jeeesuuus, you sound so tired man. Hope everything is okay.

[@tkorting]

Buy me a Red Bull and subscribe to my channel. Regards

[@rabebal9867]

is that you Severus from harry Potter talking here!!?

[@tkorting]

Please subscribe to my channel You have your mother's eyes. Regards

[@JohnVKaravitis]

5:45 You said "plus", you meant "times".

[@praveenkumarsk5957]

It looks similar to Gaussian Naive Bayes

[@tkorting]

+praveen kumar Sk thanks for your feedback. Please like and share the video and subscribe to my channel. Regards

[@kenangencol9263]

This is NOT related to MLE Estimation. Where is likelihood? Where is maximization? This video is just and no more than the discussion of multivariate gaussian distribution of data!

[@lalitagarwal63]

This is not MLE, this is bayesian approach