Collaborative Filtering : Data Science Concepts

How do recommendation engines work?

[@samuelcarlos4234]

3 weeks of classes in the university and you summarized everything in 12 min. In a way I finally could understand, ofc. Thank you for that.

[@durgaprasadvadlamoodi1271]

i agree

[@wordsexplained7565]

It's really incredible to think that out there, there are genius people sharing this type of content, and by genius, I mean someone like ritvikmath, that instead of being like the usual majority that hide their lack of knowledge behind a lot of blind nonsense formalism, gift his viewers with a deluge of knowledge like this lecture of today. We're really grateful for your work!

[@ResilientFighter]

Go Ritvik!

[@ankurdubey1441]

This is one of the best explanation of collaborative filtering on internet. Thank you !

[@legendary5320]

holy shit. an insane video, clear, concise, and to the point. Really appreciate this stellar explanation man

[@thangtran145]

To future people, do not dislike this video. It's extremely helpful! Thank you.

[@sabazainab1524]

The best youtube channel found on the internet. You are so amazing, Sir. I have watched a few other videos of yours and just clicked the subscribe button. You teach very easily and effectively. Thank you so much.

[@ritvikmath]

Wow, thanks!

[@JohnAndrewsKatadhinco]

An excellent and easy-to-understand explanation. Thanks for breaking it down and sharing some of the challenges with collaborative filtering!

[@MsKakashi2012]

Best person to explain data science concepts in the whole youtube imo

[@mr.kkquanini1156]

I just l love the approach and your way of delivering, it has really helped me a lot. Thank you

[@ResilientFighter]

Very much needed. This is extremely used in the real world, but not really much teaching in undergraduate.

[@borisdrogendijk1888]

I have noticed to be a bit late to the party with all of your videos, yet I still wanted to just let you know that you by far explain anything related to machine learning and data science out of all the guys i have stumbled upon, cheers mate!

[@ritvikmath]

Wow, thank you!

[@Somish3]

This is such a brilliant explanation, I was already pulling my hair trying to understand this concept and you just saved me, thank you! 👍

[@neotokyovid]

BRILLIANTexplanation! THX!

[@adebayoemmanuel911]

I'm liked this video to Increase the collaborative filtering so that KZbin can recommend more of this video to me.

[@mehdicharife2335]

As always, very concise and succinct explanation. Do you have other videos, or some recommendations that can help explain intuitively how matrix factorization fits into this?

[@drupad-l4i]

this was soo helpful , i was taking andrew ng courses on couresera but he didnt explain it as clearly as you were. thank you soo much.

[@ritvikmath]

Glad it was helpful!

[@bankruptLucifer]

Excellent explanation. Precise, clear and easy to follow. Thank you!

[@denjand]

Super easy to understand. Thank you so much for the great explanation!

[@mbanta]

very clear explanation that answered many questions I had from a lecture. Thanks.

[@ritvikmath]

Glad it was helpful!

[@elkanaajowi9093]

I now understand why mathematics, in itself, is a field to be studied.

[@haibhanu]

Super Explanation given to the concept. It really clarifies most of my doubts regarding the topic.. Thank You very much..

[@lvmrjb]

Watched 2 of ur videos so far that explain the concepts extremely well for a class project I have to do :) Your teaching and content are excellent!!

[@ritvikmath]

Great to hear!

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@familienolte1501]

I really liked the explanations. But is this concept superior to other cluster analysis methods such as AHC using euclidean distance for example? I mean euclidean distance is easier to understand that the cosine similarity. And for what do I need the expected rating? Wouldn't it be enough to find the most similar person and look for the highest rated film of that person, which my used person has not watched yet?

[@hounddog1]

Awesome tutorial, extremely clearly explained. Thank you!

[@ashbk1911]

Very good explanation

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@just_a_viewer5]

amazingly taught. thank you so much

[@Phil-oy2mr]

A question on cosine distance- user 1 and 3 were quite opposite on our scale and had a similarly of 0.57, so nearly 0.6. This is not very close to 0, which would indicate a true polar opposite, right? Why were 2 users here not rated near 0? What case would be? Thanks!

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@LucasElliot-k8p]

Hello, I have a question please. To get the predicted rating of User 1 for Item 4 (r1,4) Why did you multiply the similaritiy of S12 to the rating given by user 2 and S13 to the rating given by user 3. What do we call that formula or did you come up with it? What's the explanation behind it please? Also, what if the only available rating for item 4 is just the rating of user 2, can we still predict the rating that user 1 will give? Thank you so much, this is a very great video tutorial.

[@amipigeon]

Thanks for such a clear and concise explaination!

[@dragolov]

Bravo, Maestro!

[@Katharina-xt5il]

Great explanation!

[@nataliaminnigalimova7582]

Thank you for an amazing, understandable tutorial!

[@kat.simplelife]

Thank you for the precise explanation! It helps me understanding the recommendation system better. I have one question, where I'm just wondering how does latent factor fit into this?

[@omerahmaad]

Very good content thanks

[@pranjalmittal]

Really love the explanation! The video aside, I couldn't help but wonder if collaborative filtering based recommender systems that suggest content to people based on preferences of people similar to them, is one of key reasons why ideological polarity in general population is increasing on issues (such as in politics, but also beyond), because people get classified into a cluster based on similar but not the same interests, and as they see more of content in that cluster, they become even more "similar" or associative to that cluster/group they were originally somewhat but less similar to, assuming consuming content influences and creates bias in people along the lines (vector direction) of the content they consume, which I intuitively think is a fair assumption.

[@zixuanchen4506]

If U1 and U3 are polar opposite, instead of bring the score up by weighted average, can we double down if scores by U2 and U3 are far apart? something like change 0.99*2+0.57*5 into 0.99*2+0.43*1?

[@lilianaaa98]

I really love your video， thank a lot

[@stmasanti]

Geniously explained

[@gm4984]

Explained really well! Tyty :D

[@karitaalmeida1429]

Amazing explanation! Thank you very much.

[@hallvardjohnsen8302]

Great video.

[@ahmadhaziq3358]

Wow this is very well explained

[@sivaneshc109]

Amazing explanation!!! Thanks!

[@ritvikmath]

Of course!

[@eliastheod]

Excellent. Thank you very much!

[@hameddadgour]

Great content! Thank you for sharing.

[@mikeyt51386]

Great video. I just subscribed!

[@shakedg2956]

You are very talented in explaining!

[@NDBGamingHD]

great video!

[@shubhampandilwar8448]

Clearly explained. But I have some questions. Can we use users who liked(also unlike) and watched videos to recommend? How many times he has seen a particular video of a particular genre etc.(Generally, not just Netflix. ex - youtube)

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@ruslanshchuchkin]

man what a great video, thanks a lot!

[@ritvikmath]

Glad it helped!

[@adamtran5747]

love the content brother. Keep it coming.

[@scody311]

Fantastic!!!!

[@rbedson8965]

is this user based CF or item based CF, as i see the cosine is used for item based approach but again data of user is taken in user based approach. Please clear this picture for me i am new to this course

[@marathonour]

I also didn't get this point, it's great video overall, but I'm working on recomendation system and trying to figure out how svd solves this problem and should I use mult-vae instead or try content-based recsys with word2vec embeddings

[@MegaJohnwesly]

Love this explanation Ritvik.. Thank you

[@srikarsundram4633]

Do you have content on content based filtering??

[@alexivanov8800]

Excellent content! thanks!

[@ritvikmath]

Glad you liked it!

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@rmiliming]

excellently explained.

[@ayeshaayub3131]

Can you explain pearson correlation co.efficient similirity measures

[@sabazainab1524]

I wish you can add some content related to GANS as well.

[@ramonsantiago4573]

This is great, but doesn't seem to work well for spares datasets. The one thing you can do is when predicting the rating you should only divide by the sum of similarities that have ratings, otherwise your rating will be much smaller than it should be.

[@abhishekchandrashukla3814]

Sir, great work.

[@tony0731]

Well explained!

[@ritvikmath]

Thanks!

[@k5555-b4f]

Awesome videos - concepts very clearly explained

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@soner5917]

Super nice explanation. Thank you :)

[@AbhishekChandraShukla]

Is this user-based or item-based collaborative filtering?

[@leorousseau3959]

Really great video

[@xuxu4068]

Perfect. Thank you.

[@user-or7ji5hv8y]

So well explained.

[@bastianlipka2406]

love it !

[@xxshogunflames]

Awesome vid, thank you!

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@durgaprasadvadlamoodi1271]

very nice video. thanks for this.

[@MightyFrostDragon]

Thank you so much, it was super useful

[@gholamrezadar]

what if the similarities are negative??

[@ellos98]

incredible!!!

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@iaaan1245]

Bro this is awesome stuff

[@coldwinter1884]

What order would you recommend me for watching your playlists?

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@user-or7ji5hv8y]

Has another method become more popular than collaborative filtering?

[@trongnhantran3358]

Love it!!!

[@sakshamverma3114]

just love it.....

[@ccuuttww]

what is next R-SVD?

[@ccuuttww]

For alternative we can use NMF to fullfill those missing value

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.

[@levimungai1846]

please do a real project on this with actual code and explanation

[@happyhippo379]

👏

[@roykumuda]

There is clarity in your explanation. But , Is it possible for you to tune yourself into Indian accent than the american.

[@lexihilton7136]

You’re so handsome

[@hughlysds5411]

First to comment I guess

[@matthewchunk3689]

your time to shine brotha

[@venkatnetha8382]

For a 1200 long pages of question bank on real world scenarios to make you think like a data scientist. please visit: payhip.com/b/ndY6 You can download the sample pages so as to see the quality of the content.