The diagram for collaborative filtering is what seems to match with your explanation of Content Based Systems at the starting of the video. Is the explanation correct?
Thanks for this. My dataset has tv shows and i have decided to use it for my recommendation algorhitm. I am planning to convert the tv show description to vectors and then recommend based on the distance to other shows
With all due respect, you ended up explaining collaborative filtering while you intended to explain content based filtering. Content based filtering works without other user interaction. Perhaps you could add a note in the video.
In the above video, you have used Sigmoid Function to squish values into [0,1]. We know that Pearson would range the values from [-1,1]. So can we use any, or is there some logic behind choosing one of the above?
Hello Krish sir, the explanation you gave for content based filtering is actually working of collaborative filtering.....I read it on google... Is it the case?
Hi krish, nice and clear video but only one concern. How you are doing similarity of vectors that were based out of count vectorizer system which doesn't include semantic meaning?? Does this recommendation worth showing output with relevancy??
How do I make a content-based recommender with online learning? The problem with this method is, that you have to reconstruct a TFIDF matrix every time a new word needs to get added to vocabulary.
In my case the Sparse matrix formed after Tfidf is too large and when I compute cosine similarity on it, its giving me memory error. Any idea on how to solve it?
The diagram for collaborative filtering is what seems to match with your explanation of Content Based Systems at the starting of the video. Is the explanation correct?
Thanks for this. My dataset has tv shows and i have decided to use it for my recommendation algorhitm. I am planning to convert the tv show description to vectors and then recommend based on the distance to other shows
With all due respect, you ended up explaining collaborative filtering while you intended to explain content based filtering. Content based filtering works without other user interaction. Perhaps you could add a note in the video.
which algorithm is actually being used in this content based filtering 🤔???
In the above video, you have used Sigmoid Function to squish values into [0,1]. We know that Pearson would range the values from [-1,1]. So can we use any, or is there some logic behind choosing one of the above?
Hello Krish sir,
the explanation you gave for content based filtering is actually working of collaborative filtering.....I read it on google... Is it the case?
Actually yes, he explained the CF approach , I dunno why he said that it was CB
This is such great stuff! Happy I came across your channel
how can i make a fastapi of this model? I have also watched ur fastapi wala tutorial, but still have errors
Thank you so much for your explanation..
Your explanation at the start is about collaborative filtering
Yes...i too felt that
Hi krish, nice and clear video but only one concern. How you are doing similarity of vectors that were based out of count vectorizer system which doesn't include semantic meaning?? Does this recommendation worth showing output with relevancy??
How do I make a content-based recommender with online learning? The problem with this method is, that you have to reconstruct a TFIDF matrix every time a new word needs to get added to vocabulary.
Hi Krish,
I just can't seem to find your NLP playlist. Could you please help me with that?
All your videos are very clear and easy to understand, Thank you! waiting eagerly for your NLP playlist!!!!
Check in my channel
Hello sir, in which line the ml modal code is written for this recommendation system?
Thank you sir
Great video, I learned a lot. Thank you.
Sir please make videos on MATRIX FACTORISATION, NMF, SVD AND LDA FOR TOPIC MODELLING
Sir, your NLP playlist is not available to the public. Can, u please make it public. So, that we can learn the NLP concepts.
Hi @Rajat I am rewamping by NLP playlist...there also some mistakes which I want to rectify..so it will be uploaded soon
The video is very rushed, without clear explanation of why description was chosen for recommendation
Thanks Krish
In my case the Sparse matrix formed after Tfidf is too large and when I compute cosine similarity on it, its giving me memory error. Any idea on how to solve it?
Outstandingly Explained!
amazing job brother ! keep posting these kind of videos
In ln [46] of the jupyter notebook (displaying sig[0]), the first value should be 1 right (relation between the description of movie 1 and movie 1)?
Sir, Can we train our model on genres instead of overview?
can you upload content based recommendation system for e-commerce website like amazon?
Hi, if this method is applied on a larger dataset, "nnz of the result is too large" will occur. Any workaround for this? Thanks
I am also facing same problem. if any solution please share. thank you
Good job sir
I got an error list object is not callable
sigmoid_kernel is really heavy process , is there any better way ?
How can I check it for a new data point??
Amazing
can you please explain matrix factorization(SVD)?
How could i get the code for this project
thank you sir!
May I get the code.
github.com/krishnaik06/Recommendation_complete_tutorial/blob/master/Content%20Based%20Recommendation%20Engines%20using%20Python.ipynb
please , provide code link
sir please share source code
Gui please
Does anyone know which loss function we use for recommender system?
Square loss function