- 1
- 182 203
Andrius Knispelis
Denmark
Приєднався 17 тра 2015
LDA Topic Models
LDA Topic Models is a powerful tool for extracting meaning from text. In this video I talk about the idea behind the LDA itself, why does it work, what are the free tools and frameworks that can be used, what LDA parameters are tuneable, what do they mean in terms of your specific use case and what to look for when you evaluate it.
Переглядів: 182 208
This video topic on LDA modeling is the best layman's presentation thus far, but even more so applicable in the recent state of AI modeling. Thank you dearly.
thank you so much, happy you liked it!
Thanks a lot
Thank you!!! Great help!!!
This is great, clear, and practical. Excellent package in 21 minutes. Thank you!
Thank you, very useful
amazing !!!!
That's a pretty great video for LDA! Thank you sir for making this. I just have a simple question. Given a magazine, can we name its topic using the words that have high prob in the highest prob topic associated to this magazine?
Thank you soooooooooooooo much. The best topic modelling presentation ever!!!!!! Many thanks from a linguistic student.
One of the best presentations I've ever seen .. Thanks
Fantastic! Thank you!
Iiiiiii
could you share us another new video how topic models applying to real world problems or solutions
Really clear and concise way to explain this complex topic. Thanks
can we apply LDA modeling on images (not text image).
Instantly can see, when someone is talking about a topic in which he was involved himself. Great job! You definitely know what you are talking about. “If you can't explain it to a six year old, you don't understand it yourself.” ― Albert Einstein.
This one was super helpful, thank you very much!
Compact, crisp and strong narrative video presentation...I watched it only 2 times and understand the process thoroughly....1 question, just to get your insight..Is LDA can be combined with systematic literature review protocol (SLR) and the produced model in LDA is similar with structural equation modeling (SEM) model?
wow ,this is a great video!! and i like the color scheme you used, could you share it or where i can download the similar styles? thanks!
hi Todd, happy to hear you liked it :) for colors i used one of the styles that i found here: flatuicolors.com (my go to place for color schemes). For a font i used "Helvetica Neue".
Great video! One question - what is a magazine?
A question for future historians! :) Magazine is 40% articles of a certain, usually shared, topic, and 60% adds. I know, i didnt belive it either, but 20th century was a wild one in terms of ideas like this
great job. but why didn't you describe the prior distribution this model uses?
Thanks Zahra, happy you liked it! I didnt go into priors since i didnt touch them - i left them as their default values in gensim. I mostly experimented preprocessing of text corpus, number of topics and interpretation of the results :)
one of the best presentation for beginner like me
Amzing presentation and explanation
Amazing!!!
This presentation is so much informative. Thanks!
thank you so much for making such a complex concept relatively easy to comprehend
Make more content!
This is really great! Love the excellent visualization and methodical explanation.
amazing explanation!
This is a lot of help! Many thanks~
Man I loved this video. It helped me so much!! Really apreciated it. Now my master degree is on the right track once again!!!
The best presentation I’ve seen for LDA’s and most other themes. Outstanding work thank you for producing it!
Found this trying to learn about linear discriminant analysis, stayed because it was a week put together presentation on an interesting topic
it was as help full as possible thank you so much
A slightly different question. Which tool was used to create this presentation?
THANK YOU! :) Everything was done in Keynote (mac version of a Powerpoint). It comes with all the animations you see, and it lets you record sound as well and then export everything straight into video.
Wow! Blown over by the video. It was easy to follow and gained a lot of information for implementing my model. Thank you
This presentation is excellent. Thanks.
Thanks a lot sir
Amazing intro to LDA, thank you very much
Dude, such a great presentation. Thank you very much for this superb explanation! :)
Very well explained. Did you use Mallet along with Gensim or was it only Gensim?
Important question: I build a model like this on my own using simple SVD and vectorizing my documents. At the moment I have around 10k documents and searching for the closet 10 documents or so is quite fast. (~0.001s). However, at this moment, I am just brute-forcing getting the closest vector in linear time and if my database will be bigger (over 1M or even 10M documents) I need some kind of indexing to make the search much faster. Do you have any suggestions? I feel confident to design an approximate algorithm by myself, but I would prefer an exact solution running in log(n) time.
what i did was use K means to cluster the whole space of documents. Then the number of clusters will depend on number of themes (not the documents). Each cluster is then represented by a single LDA distribution, and then the search was split into two spets: 1) find a cluster, 2) find nearest neighbors in that cluster That helped us a lot. Well, that and the fact that while LDA model was trained and processed in Python, we used faster and more efficient languages for finding similarities
Hey Andrius, great job explaining the topic modeling concept and relating it with the use-case of magazines. I really enjoyed the presentation, the graphics, video, and the entire layout plan. Kudos. Once again, thank you for posting it here.
Thank you for that beautiful presentation! I learnt a lot from it and enjoyed it immensely
Amazing explanation! Thank you for this.
Wow. Fantastic explanation. Thanks so much.
wow just wow dude i like ur explanation very much
best explanation
What an amazing presentation Andrius. Very well explained and nicely crafted. Clearly demonstrates your deep understanding of LDA and fantastic communication skills!
Thank you Alok, glad to know you liked it!
[Remove if a word appears in more than 10% of the articles. Remove if a word appears in less than 20 articles.] Can someone explain these actions in detail? Why do we need this?
This is the part when dictionary is built, it's about removing words that appear too many times or too few times. Check out the tutorials here radimrehurek.com/gensim/auto_examples/index.html Hope it helps :) cheers!
OMG literally every single word you said in this video is super helpful to me. Thanks a ton!!
Thanks Peeta! :) Happy you liked it!