Just a suggestion: for a better understanding of the loaded 'data' and 'movies' and for debugging, it is useful to include: " print(data.head()) " and print("movies.head()). Be Sure to comment out such print statements and also be cautious of comments within the code since 'streamlit' often complains about any code that interferes with its startup or running.
P.S. As you know, good practice -- under MacOS, MS Win or Linux, is to either use Anaconda ('mini-conda') to isolate you dev projects, or, at least 'venv'.
Although, 'cosine_similarity' is commonly used for matching query to vector_db, for a "Recommender System" is there another function that you might recommend ?
Thank You Very Much for Your Quick Replies. Looking forward to your next Vids !!!!
Thank You. Very Useful and Timely. NEW SUBSCRIBER ! Also, I am working on a vector representation of User's preferences, likes and dislikes ...
Thank you!
Just a suggestion: for a better understanding of the loaded 'data' and 'movies' and for debugging, it is useful to include: " print(data.head()) " and print("movies.head()). Be Sure to comment out such print statements and also be cautious of comments within the code since 'streamlit' often complains about any code that interferes with its startup or running.
thanks for suggestion! Will keep in mind
source codes: github.com/GoingLight/movie-recommender
P.S. As you know, good practice -- under MacOS, MS Win or Linux, is to either use Anaconda ('mini-conda') to isolate you dev projects, or, at least 'venv'.
yes, you are right, I use conda as well, and it is definetly more convinient
Although, 'cosine_similarity' is commonly used for matching query to vector_db, for a "Recommender System" is there another function that you might recommend ?
another function you might consider is dot product similarity, useful when you also want to consider the magnitude