got it sir 1- copy ur location followed by the name of the file , it will be in the properties area 2-the extension will be given in the properties column check and enter
Easy and Simple video to understand Recommendation System. Thanks for all the efforts. Doubt: 1. Any intuition behind filtering #ratings > 100, I mean why 100? 2. Can we apply filter on correlation and sort why #ratings? if yes, how to choose correct value for correlation? 3. Does the correlation only consider wether rating is given by similar user (Yes/No) or it also consider what rating is given (0-5). Bcz lets say User A watched and rated Movie b -> 2.5 rating, Movie C -> 1.5 rating. This indicates that user A watched these movies but do not like them. So if correlation only consider wether rating is given or not in that case user X might get recommendations (similar to User A i.e Movie B,C) and he may not like them.
Hi Krish First of all thank you for all the efforts you are putting into the Data Science community. I would like to know how can we get the book which you have authored in regards to Practical Implementation of Python in finance.
It's so that the recommender can be more accurate. If you have more data on a particular movie, there's a greater chance that the movie that is being recommended will actually be a good recommendation. For example, I would be more inclined to watch 1 movie that has been recommended by 100 people rather than a movie that has been recommended by 1.
Really nice tutorial, you got my sub! i have 2 questions: Did you use a "content basic filtering" or a "collaborative filtering"? Why do not use a non negative matrix factorization?
It is collaborative filtering. In collaborative filtering, it is the memory-based approach in which Krish sir applied the item-item method which can be seen when they are finding similar movies (or say the best correlation between movies) like Star-wars.
Very good explanation and I have a doubt that we are saying that this predicts and recommends now after this algorithm implementation can we predict movie and recommend to the new user? can you please answer this question??
hello Krish sir your tutorials are awesome, very informative, m getting a doubt in this project while uploading a file not found error (u.data) pls do needfull
But sir instead of two user want to find out recommendations for each movie then how should i proceed ,cause creating correlation data frame is not possible for all of this individual
Thank you for the tutorial. I have a question though. What is the intuition behind using correlation of ratings? Correlation would only give information about which other movie was proportionally rated compared to a given movie. Do you have a video where you explain the intuition behind why this works?
The correlation of ratings helps you understand the relationship between the movies. Since the matrix is based on individual user(s) ratings, thus we can easily correlate his likings against all the similar rated movies. In short, if user A likes movies B, and C and then a similar user to A might also like movies B, and C.
It is just an easy and basic way to understand Recommendation Systems, however, in real life, these algorithms do not work or are not preferred. You need to understand the correlation matrix as it is used as a base model to understand the business requirements and to architect a solution for the same.
Join function is not working for my data and error is generated that column overlap but no suffix specified please help in solving this and the video is too good
Hi Krish i learned a lot from this video but i have a doubt "How did we find correlation of Star Wars movie with other movies" i am confused. Please help me on this.
Thank you for the tutorial. When I replace the movie name with "Shawshank Redemption, The (1994)" I get only NaN's in the similarity matrix, resulting in empty ratings dataframe. What could be the problem? is it that there is no similarity between this movie and other movies? Is there a sure way to get around this problem? There are also a few movies that repeat this behavior too. "Shawshank Redemption, The (1994)" is just an example. thank you
The recommendation here is based on ratings only given by users not attributes of the movie , it has to extended to attributes of the movie (genre play time etc) to solve the Nan issue . Note This is a starting guide to recommendation algos and a great one at that . thanks @krish naik
@@maybe9357 maybe it will be done by extracting pin point data such as genre through api or if you get a dataset in which there is genre column you could do this
Does this mean, if someone has watched a 1 rated movie. Next time similar 1 rated movies would be recommended to him which has more than 100 ratings as 1 rating ? If that is the case , it is not the ideal case...right ?
Hello Sir, I am getting following error while trying to load movies csv file "'utf-8' codec can't decode byte 0xe9 in position 31: invalid continuation byte". I will be grateful if could help me out.
Sir can you make a video on fitness based recommendation system. and can you tell me how to collect data set of different people and recommend them food and exercise according to that data set. plz reply sir....
got it sir
1- copy ur location followed by the name of the file , it will be in the properties area
2-the extension will be given in the properties column check and enter
Krish sir. Cant thank you more. Such a hardworking , insightful and talented person changing lives of people.
Excellent Krish. Your explanation is right on the money and crystal-clear. Thank you so much. Kindly continue with all the good work.....
Thanks Jeff.
Sir you are such an amazing teacher , i understood every single bit of what you have explained thankyou very much sir.
Easy and Simple video to understand Recommendation System. Thanks for all the efforts.
Doubt:
1. Any intuition behind filtering #ratings > 100, I mean why 100?
2. Can we apply filter on correlation and sort why #ratings? if yes, how to choose correct value for correlation?
3. Does the correlation only consider wether rating is given by similar user (Yes/No) or it also consider what rating is given (0-5).
Bcz lets say User A watched and rated Movie b -> 2.5 rating, Movie C -> 1.5 rating.
This indicates that user A watched these movies but do not like them.
So if correlation only consider wether rating is given or not in that case user X might get recommendations (similar to User A i.e Movie B,C) and he may not like them.
Good explained. Sir example you explained in last video Netflix was collaborative and contents based is for website. In this video is different
This one is based on correlation
Brother i cant found u.data plz send me url
Very well simply explained , i love the way you teach . Looking forward to join the online ml class your's.
Nice tutorial on Recommender System.
Very good Krish Naik. Really like it. I hope you continue with this series. Got my sub. Best regards
Yes definitely
This is an excellent starting guide to recommendation systems.Nice work @krish naik
simple, crisp and helpful
excellenet ..crisp and to the point ..great video
Thank you very much ! You saved my life !!
now say bro .. he saved then ?
Hi Krish
First of all thank you for all the efforts you are putting into the Data Science community.
I would like to know how can we get the book which you have authored in regards to Practical Implementation of Python in finance.
The book is out of stock now
@@krishnaik06 is the book in stock now?
Nice tutorial. But I have a quick question, Why you have used no. Of rating >100 as filtering criteria ?
It's so that the recommender can be more accurate. If you have more data on a particular movie, there's a greater chance that the movie that is being recommended will actually be a good recommendation. For example, I would be more inclined to watch 1 movie that has been recommended by 100 people rather than a movie that has been recommended by 1.
Really nice tutorial, you got my sub!
i have 2 questions: Did you use a "content basic filtering" or a "collaborative filtering"? Why do not use a non negative matrix factorization?
I also have the same question
Just a beginner...so anyone can help regarding this.
it could't read movie id titles in dataset .i'm using in colab .
It is collaborative filtering. In collaborative filtering, it is the memory-based approach in which Krish sir applied the item-item method which can be seen when they are finding similar movies (or say the best correlation between movies) like Star-wars.
Amazing tutorial krish sir helped alot thank you
Hey @krish Naik, can you tell me why have you taken bin=70? Nice video though, thanks a lot.
Great job 👍. Can you create a vedio on recommendation system using autoencoder or RBM
Very good explanation and I have a doubt that we are saying that this predicts and recommends now after this algorithm implementation can we predict movie and recommend to the new user? can you please answer this question??
Great tutorial. It helped me a lot to understand my project
How to find recommended products accuracy
Good Job and very easy to understand .
Can u give the csv file of this dataset?
Helping Hand for Beginners!! Thank you for uploading this video. Hope to see more videos on Data science projects.!
Thanks. Please subscribe the channel and press the bell icon, so that you get the notification as soon I upload a video
hello Krish sir your tutorials are awesome, very informative, m getting a doubt in this project while uploading a file not found error (u.data) pls do needfull
Great video sir. One question- Which algorithm are you using or wld say is best for building a recommendation system??
sir can you teach how to import dataset which is not in the form of csv in this video how did you import this data to jupyter notebook
Great video thank you so much for this amazing tutorial
But sir instead of two user want to find out recommendations for each movie then how should i proceed ,cause creating correlation data frame is not possible for all of this individual
Thank you for the tutorial. I have a question though. What is the intuition behind using correlation of ratings? Correlation would only give information about which other movie was proportionally rated compared to a given movie. Do you have a video where you explain the intuition behind why this works?
The correlation of ratings helps you understand the relationship between the movies. Since the matrix is based on individual user(s) ratings, thus we can easily correlate his likings against all the similar rated movies. In short, if user A likes movies B, and C and then a similar user to A might also like movies B, and C.
It is just an easy and basic way to understand Recommendation Systems, however, in real life, these algorithms do not work or are not preferred. You need to understand the correlation matrix as it is used as a base model to understand the business requirements and to architect a solution for the same.
Nice lectures Krish, Can you please explain one real time project from the scratch how to do? how to get the data?steps to be followed etc
Sir can you please suggest few problem statements on Recommender System
Thanks Krish
@Krish Naik what is the difference between collaborative filtering and Market basket analysis?
You re the best
Join function is not working for my data and error is generated that column overlap but no suffix specified please help in solving this and the video is too good
Hi Krish i learned a lot from this video but i have a doubt "How did we find correlation of Star Wars movie with other movies" i am confused. Please help me on this.
Thanks for awesome content...
This video helped a lot. Thank you!
Very Nice Video
Great Job
Really useful video - thank you very much!
good tutorial sir
super explanation! thanks
really well explained!!
Thank you for the tutorial. When I replace the movie name with "Shawshank Redemption, The (1994)" I get only NaN's in the similarity matrix, resulting in empty ratings dataframe. What could be the problem? is it that there is no similarity between this movie and other movies? Is there a sure way to get around this problem? There are also a few movies that repeat this behavior too. "Shawshank Redemption, The (1994)" is just an example. thank you
The recommendation here is based on ratings only given by users not attributes of the movie , it has to extended to attributes of the movie (genre play time etc) to solve the Nan issue .
Note This is a starting guide to recommendation algos and a great one at that . thanks @krish naik
@@subirpaul9472 how do we append that?
@@maybe9357 maybe it will be done by extracting pin point data such as genre through api or if you get a dataset in which there is genre column you could do this
Perfect sir. This will help me soo much
I have download the data but my data doesn't get import on jupyter in giving error.
Fab !!!!!!!!!!!!!!!!!!!!!!!!!
sir ab isko pycharm sa kaisa connect kare plzz bta dijiye
i learn so much Thx !
Does this mean, if someone has watched a 1 rated movie. Next time similar 1 rated movies would be recommended to him which has more than 100 ratings as 1 rating ? If that is the case , it is not the ideal case...right ?
will this example comes under collaborative or content-based filtering
Collab
Sir do u have any journal/research paper on this project??
Very great content, thanks
Amazing work bro, really simple but gives alot of information
What is the logic behind correlation in Movie Recommender system
First know the correlation definition... You can get to know the answer
THANK YOU
Can you please make a tutorial on how to deploy this on an web app please
Great tutorial especially for beginners.
Please tell me uniqueness of this project ??
My teacher is asking me ....please tell me uniqueness of this project
Why can't you take the product of the number of ratings and the ratings ?
This was really nice thank you
Can't seem to pivot my dataframe since I have too many rows. Any tips to do this?
Hi Krish. How can we find Accuracy between two movies? How can we do comparative analysis between two movies?
what do you mean by "Comparative analysis", can you explain?
You can use estimation tools such as RMSE
Thanks a lot... Very good explanation
What is the algorithm used in this project???
If the user has less than 2 entries, how do we remove them from the list?
If the ratings in data is from negative to positive value so can I do normalization?
is it possible to include an evaluation metrics?
So is this content based filtering or collaborative filtering?
I watch ur work and try to do it with ur same algo but different datasets and now im stock in merge datasets, can u help me fix my error?
Sir I can't download the u.data file.and I don't know what is the problem
Like it buddy.
Are you using Pearson's correlation here ?
Thanks!
How to add this recommendation result in software?
Why python is only used for this purpose? why not any other language?
please reply!
Thank you!
Sir daraset kaha se lu kaise aur u.data ky hai plz help
which command is used in project
Nice tutorial Krish.
Could you also let me know which software are you using to capture the screen?
Is this collaborative filtering?
Hlo.. sir what u mean by time stamp in this proj
Thanks alot !! awesome explanation.
Unfortunately, doesn't work if you have 100 000 books and 150 000 users) but EDA is very helpful
Hi, can we create a UI for this project?? If yes, Can you say how to do?
Hello sir just wanted to know is this related to machine learning or AI?
so how do we evaluate this model?
Can I use ur code for doing this video in my language .. I will give you credits...
Post the link of article you followed from 'Towards data science' please!
this one is the link
Please help with the 'u.data' file as it's showing error : File u.data does not exist: 'u.data' ?? downloaded the file in the same directory.
I think you forget the '.txt' extension (pd.read_csv('u.data.txt',sep='\t'))
@@shelaraarti6082 even showing same error plz help me out bruh
How did you evaluate your model?
Hello Sir, I am getting following error while trying to load movies csv file "'utf-8' codec can't decode byte 0xe9 in position 31: invalid continuation byte". I will be grateful if could help me out.
Write encoding ='utf-8
this is the udemy copied data or file
named: perion data
sir could you please provide the link of file of u.data .
m not able to download the file u.data.
please help me out
sir problem is sorted out
thanks for such wonderful video
Brother could u plz tell me how it solve
@@walkWithDinkar can u send me url of u .data dataset its help me a lot
Sir can you make a video on fitness based recommendation system.
and can you tell me how to collect data set of different people and recommend them food and exercise according to that data set.
plz reply sir....
what is the accuracy?
what if we have multiple ratings.... rating on differnt criterias?
It's based on use case , you can make a recommendation engine based on each rating type or take a mean
sir please share link of dataset