KeyBERT Keyword, KeyPhrase extraction using BERT embeddings

[@sajansudhir3317]

Thanks for great tutorial packages.Your lecture about SpaCy transformers on Text Classification and Custom Named Entity Recognition helped me to impress interviewer and because of that right now I am officially with startup IT company.Pls be on your course and introduce us latest technology to cope up with AI domain.....

[@RitheshSreenivasan]

Congratulations!! Happy to know that my videos has helped you. I will try to be consistent with my videos

[@gulabpatel1480]

love and appreciate the hard work you put to create such a great videos...
a kind request, if possible plz try to mention custom training as well even just a high level overview is enough at the end of video....Thank you

[@RitheshSreenivasan]

Thank You!! I would mention custom training where appropriate.

[@karthikgali]

Thanks for the great introduction to KeyBERT. I have a question about handling new keywords (vocabulary). For example: COVID was never seen before 2020 and is the one of the important Keyword - how would BERT based models work on OOV words.

[@RitheshSreenivasan]

Bert can use sub word tokeniser. So if the OOV word can be split into subwords which may be present in vocabulary you may get some results.OOV words are always a challenge to any model and it depends on how OOV words are handled

[@karthikgali]

Agree with you. Thanks for your inputs.

[@jeromesoh9249]

Hey, was wondering if i could do the same for multiple customer reviews instead of just one?

[@RitheshSreenivasan]

I did this video some time back. Please refer to KeyBERT documentation for the same. At that time it was extracting keywords from a single document. Probably you can loop over a set of reviews

[@jaivalani4609]

Thanks Rithesh , Can this model be used to find Top 5 issue or top 5 appreciations reported by customers. Say 1000 customers have given a feedback about my product
Also how would i interpret these prob given with differrent options

[@RitheshSreenivasan]

I have made this video last year so please have a look at the paper.

[@kaushikjoshi8191]

Can you please provide link or name of that video?

[@bhagirathradadiya9753]

Thank you for this tutorial. I have one query about language support. Is keybert support French or Spanish text?

[@RitheshSreenivasan]

Please refer to the GitHub page of KeyBERT

[@Zeeshanmys]

Sreeni.. Do we have option to extract key value pair information from raw text using NLP..??

[@RitheshSreenivasan]

What do you mean by a key value pair and how do they occur in text. Depending on how they occur you have different ways to extract them using NLP

[@sanjaysan9249]

I have one query about language support. Is keybert support Kannada Language? sir

[@RitheshSreenivasan]

I don’t think so

[@adityancg5146]

No matter how you change the parameters, does it only give 5 key phrases?

[@RitheshSreenivasan]

Have a look at the documentation

[@ashishsaman]

There’s a parameter for it. I think it’s called ‘n_words’ or something like that.

[@Munna_9652]

ImportError: cannot import name 'keyBERT' from 'keybert' (/usr/local/lib/python3.7/dist-packages/keybert/__init__.py)
hii im getting error like this idk why
please help me to find the solution

[@RitheshSreenivasan]

Looks like some library install issue

[@Munna_9652]

@@RitheshSreenivasan no i saw every step from youtube video i didnt miss anything ..that works for that guy but not for me

[@RitheshSreenivasan]

Google and you may find a solution

[@Munna_9652]

@@RitheshSreenivasan ohh i didnt get that idea yaaarr 😒😒😒
1st thing i did that only

[@RitheshSreenivasan]

If google can’t give you a solution how can I give. You can raise an issue on KeyBERT GitHub page and may be they can give you a solution