Hi, while you are trying not to present the recommendation of 'toy story' itself, by adding: print(df_movies['title'][i].where(i!=idx)) you still have that in the first line as NaN
Nice explanation, could you please suggest a way to check the accuracy of this model? I implemented a similar model, but do not get how to show model accuracy/RMSE.
Sir, how to build recommendation system based on dataset that has no numerical values. I mean the dataset contains [user_id, post_id, post_type], so that using this dataset we have to recommend similar post for the user
Hey Guru Raja, that's an interesting question, such recommendations are usually made by content-based filtering which means you need a way to calculate similarity of the selected post(text) with all the other posts, one method used in such scenarios is to compute the term-frequency vectors of the posts and obtain the cosine similarity of those vectors, based on which recommendations can be made, hopefully this helps !
It couldn't be more straightforward and easier but with your explanation, great work!
Nice video. my concern is how do we check the accuracy/RMSE
Hi, while you are trying not to present the recommendation of 'toy story' itself, by adding: print(df_movies['title'][i].where(i!=idx))
you still have that in the first line as NaN
while running pivot, i am getting index error..how to solve that
Nice explanation, could you please suggest a way to check the accuracy of this model? I implemented a similar model, but do not get how to show model accuracy/RMSE.
did you find a way?
sir how to build user based collaborative recommendation using knn?
What algorithm do you use?
hey thanks for the great video!
can you please help me out with how to display the ratings of the movies along with the names??
Nice video, i would like to predict the ratings of the movies as well, how can i proceed ?
hello sir, is this item-item collaborative filtering if yes then can you make a video about user-user filtering?
Hello sir?
Is this done by euclidean distance or by cosine similarity?
cosine
Could someone tell me how to save this model in pickle or joblib
Do you make paid projects too ?
Very well explained!
Sir, how to build recommendation system based on dataset that has no numerical values. I mean the dataset contains [user_id, post_id, post_type], so that using this dataset we have to recommend similar post for the user
Hey Guru Raja, that's an interesting question, such recommendations are usually made by content-based filtering which means you need a way to calculate similarity of the selected post(text) with all the other posts, one method used in such scenarios is to compute the term-frequency vectors of the posts and obtain the cosine similarity of those vectors, based on which recommendations can be made, hopefully this helps !
Thank you sir, for your response. I will definitely recommend this channel to all my friends.🙂
Have you been able to build this. I am having the same issues and I don't know how to go about it.
Is there any procedure to evaluate this model?
did you find any way to evaluate it?
Nope
sir how to install fuzzywuzzy
pip install fuzzywuzzy
Great Video Brother :))
Thanks alot brother, stay tuned for more :)
By the Ritik here,
We met few days ago in NED after the event
Bro your content and your accent are so awesome.
Great keep it up
I'm getting a key error here