Intuition behind Latent Dirichlet Allocation (LDA) for Topic Modeling

Video quality	The size	Download

Information Intuition behind Latent Dirichlet Allocation (LDA) for Topic Modeling

Title	:	Intuition behind Latent Dirichlet Allocation (LDA) for Topic Modeling
Lasting	:	6.01
Date of publication	:
Views	:	142 rb

how can random assignment of topics lead to correct assignment of topics?
Comment from : @preethamkumar91

Great videobrbrJust a quick correction, in @1:11 you have written that "Stemming: merging words that are equivalent in meaning" That actually is called LemmatizationbrbrStemming actually reduces the words to their radicals, since the topic analysis can be done with just them, reducing the size of the analyzed data
Comment from : @VictorAlmeida27

Well explainedsimple and clear! Thank you!
Comment from : @mariarocque7384

Useful Thank you!
Comment from : @Juan-Hdez

1000th like
Comment from : @parthicle

Observations are referred to as words The feature set is referred to as vocabulary A feature is referred to as a document And the resulting categories are referred to as topicsbrbrIs this correct understanding?brplease correct me if I am wrong
Comment from : @rajaramwalavalkar9187

ur way of speaking gives me a gamer vibe bro
Comment from : @Admin_REX

Great explanation! What type of inference is this? (Gibbs sampling, variational Bayes, etc)
Comment from : @michaelyoder7250

always indian man always
Comment from : @hence0182

this is not clear why hes subtracting it 3:10
Comment from : @monamyers8324

What's the significance of the word 'dynamic' in the slide VII?
Comment from : @krishnagarg6870

woow woow it was very helpful, as I am working on this algo
Comment from : @_xkim00

Thank you so much Great video you used such a simple example to explain this LDA, wish I saw your video long time ago thanks again
Comment from : @robindong3802

Man, I lost you mid-way in 'topics' but anyway I could get an idea what LDA actually is Thanks :)
Comment from : @akshaygera9097

Under this algorithm, a word would never be reassigned to the same topic again?
Comment from : @abcd12272

How was "World Cup" reassigned to topic 1 based on the multiplication of the 2 matrices?
Comment from : @abcd12272

Thank you for the cleanest and simplest explanation
Comment from : @shagshaq

Sir I had done LDA using Scikitlearn library brWhen should we use Gensim Library or anything is same
Comment from : @consistentthoughts826

Slide no 8 please
Comment from : @ravindarmadishetty736

Very confusing
Comment from : @snandi1603

You explained really well But try to elaborate the explanation so that it can be understood in one go Going over the video once again is cumbersome
Comment from : @parthaprateempatra4278

Thank you for this video! Clearly explained I would request you for an video on how to perform dirichlet regression using R or python Thank you
Comment from : @priyanatraj5634

Hi Bhavesh, can you please explain how the area rather the probabilities are calculated
Comment from : @sudeshnadutta5702

You started well, with good examples but at mid and at end, it was difficult for a newbie of this field like me to understand
Comment from : @romy5994

what's the corpus argument passed ?
Comment from : @vibewithalexa

Thank you so much Precise and clear explanation !!
Comment from : @rajsinghmaan3095

couldn't understand- how much doc like topic* how much topic like word!!
Comment from : @arpitqw1

Please make a video on KL divergence, it will be a great help brregards
Comment from : @ash_engineering

Best explaination ever continue !!!
Comment from : @fitnesscoach7

clear and crisp
Comment from : @SajeedSk

the matrix improvement part could be explained better but definitely the best video on the topic(no pun intended) Thanks
Comment from : @VishalSingh-dl8oy

really you are the best regards
Comment from : @Aliabbashassan3402

Quick question regarding 2:32, you point out that a word is associated with multiple topics? I thought a word can only be associated with a one topic, while a document can be associated with multiple topics
Comment from : @MasayoMusic

So one iteration of the algorithm is the same as going through the document and reassigning the topic for each word of the document, and do that for all the documents? Would it be wrong if I did the iteration N times on a single doc and did those N iterations for each document? Does the order of operations matter here?
Comment from : @fancypants7533

Tks Bhavesh, good work!
Comment from : @TheEscolaris

This is so helpful Thanks Bhavesh :)
Comment from : @rohanchadha3506

hey thats a good work out there and can you please give a link or something for the presentation i will bw really helpful
Comment from : @shubhammishra6687

This is by far the best explanation of LDA I went through literally dozens of videos and none of them explained the technical details Thank you for this video
Comment from : @nishantjha6412

Wonderfully explained, I was reluctant to read the Andre Ng paper's theoretical and mathematical explanation on LDA, and this gave me the whole idea in just one go Great work!
Comment from : @100damen