For the people who are getting confused between cosine similarity and cosine distance. In cosine similarity if the value is towards 1 then it would be considered as very similar. If the value is towards 0 then there is no or less similarity between the two points or vectors. In the video at 26:50 Krish sir is talking about cosine distance not cosine similarity. So in cosine distance if the value is more towards 0 then there is a similarity between the points or vectors and vice versa.
Yes, you're correct! Cosine Similarity: Measures similarity, ranging from -1 (dissimilar) to 1 (similar). A higher value means more similarity in direction. Cosine Distance: Although not a standard term, it is commonly used to denote 1−cosine_similarity. It quantifies dissimilarity, where a lower value (closer to 0) means more similarity and a higher value means more dissimilarity.
I really appreciate ur efforts ! and whenever am in doubt of ML concepts, the first thing comes to my mind is to search through ur channel ! I really love ur teaching ! and idk how many times i have commented this in the comment section !!!
Hi Krish, I'm super grateful for your videos. You are helping a lot of students who cannot afford to pay and learn. I got into DS because of you. Thank you for everything ! please continue to do whatever you are doing.
Vocabulary in one hot encoding is usually a set, so the size of the vocabulary in question no 2 is 3, not 4. One hot encoding is a single 1 with rest as zeros.
Thank you Krish, really grateful for the video. I have learnt almost everything from your videos. However sometimes it is really difficult to understand the context between the lines and sometimes you say self contradictory things. Similarly here it was difficult to understand the use of O/P feature (Is, related, to ) and the o/p layer of ANN which again without context again comes down to the words in document. please try to elaborate in a sequential model. Maybe its just me who is facing the issue to understand and maybe this comment will get ignore all together. But thanks for the efforts. I will try to browse word2vec for better understanding. I have gone through each community session of ML and deep learning and now I am on NLP. I am following the roadmap you have suggested. I am also in between a career transition from core disease and biology research to data science. Hopefully I will be able to make a successful transition. Keep the good word going for people like me...
Krish,, Great video again, thank you so much. Just one question: what hardware are you using to write with your notes in these videos? Appreciate your answer.
46:46 So you mean If we want to represent a word with 300 dimensions then window_size has to be 300 ? Because in your case window_size is 5 and you represent word by 5 dimension vector
One simple gratitude:thanku very much ..I cleared so many doubts .just one qsn if I will do codemix sentiment analysis how to consider Hindi words like I don't want to use Google translator like if the sentences:aaj to fun day tha..how I will input this via python??
Hi Sir, I am getting the follwoing error while importing gensim library: *ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject* Please help me to solve this error
For the people who are getting confused between cosine similarity and cosine distance. In cosine similarity if the value is towards 1 then it would be considered as very similar. If the value is towards 0 then there is no or less similarity between the two points or vectors. In the video at 26:50 Krish sir is talking about cosine distance not cosine similarity. So in cosine distance if the value is more towards 0 then there is a similarity between the points or vectors and vice versa.
thx man!
Yes, you're correct!
Cosine Similarity: Measures similarity, ranging from -1 (dissimilar) to 1 (similar). A higher value means more similarity in direction.
Cosine Distance: Although not a standard term, it is commonly used to denote 1−cosine_similarity. It quantifies dissimilarity, where a lower value (closer to 0) means more similarity and a higher value means more dissimilarity.
I'm really loving the content. Thank you for your time and patience.
You are one of the best teacher here. Whatever you teach, goes through so easily. I appreciate your dedication. Thumbs up from Pakistan
I really appreciate ur efforts ! and whenever am in doubt of ML concepts, the first thing comes to my mind is to search through ur channel ! I really love ur teaching ! and idk how many times i have commented this in the comment section !!!
I don't know how to thank you man .. you are the best
Hi Krish, I'm super grateful for your videos. You are helping a lot of students who cannot afford to pay and learn. I got into DS because of you. Thank you for everything ! please continue to do whatever you are doing.
Sir after nlp start community session of computer vision complete series
Vocabulary in one hot encoding is usually a set, so the size of the vocabulary in question no 2 is 3, not 4. One hot encoding is a single 1 with rest as zeros.
Thank you Krish, really grateful for the video. I have learnt almost everything from your videos. However sometimes it is really difficult to understand the context between the lines and sometimes you say self contradictory things. Similarly here it was difficult to understand the use of O/P feature (Is, related, to ) and the o/p layer of ANN which again without context again comes down to the words in document. please try to elaborate in a sequential model. Maybe its just me who is facing the issue to understand and maybe this comment will get ignore all together. But thanks for the efforts. I will try to browse word2vec for better understanding. I have gone through each community session of ML and deep learning and now I am on NLP. I am following the roadmap you have suggested. I am also in between a career transition from core disease and biology research to data science. Hopefully I will be able to make a successful transition. Keep the good word going for people like me...
Thank you for your time and patience ❤
Krish,, Great video again, thank you so much. Just one question: what hardware are you using to write with your notes in these videos? Appreciate your answer.
Thanks a lot for this amazing lecture
Thanks!
Hey I think that is one hot encoding you have shown for CBOW it's not bag of words
thank you for creating such value contents
@1:16:52 ,word embeeding doesnt have 500 dimensions for 500 words if thats the case then there is no difference between BOW and Word embeedings
@krish, @36:41 method you used to represent the sentence is One-Hot encoding not BOW
Hard to understand totally without examples but thank you to the effort
The best channel for learning Data Science. Thank you for all your effort and knowledge sharing!
46:46 So you mean If we want to represent a word with 300 dimensions then window_size has to be 300 ? Because in your case window_size is 5 and you represent word by 5 dimension vector
Thanks a lot sir.
One simple gratitude:thanku very much ..I cleared so many doubts .just one qsn if I will do codemix sentiment analysis how to consider Hindi words like I don't want to use Google translator like if the sentences:aaj to fun day tha..how I will input this via python??
Keep teaching NLP sir!!
I am unable to see your notes on the dashboard. Could you please tell me the exact location where I can find those?
But noun and preposition conjunction will have high frequency
One hot encoding if you can repeat please
where is the google colab link.Can you please share.I am unable to find it
'IS' is a stop word right.. we will remove it in the text pre processing stage.. then why are you considering 'is' as your output word ?
Dashboard link is not working. Could you please check it? Thanks in Advance.
Great
Super session
can anyone help me with the ipynb file in this video (please provide the link)?
kindly share notebook file the link you provide removed form website or may have an issue
sir i need the study metarial but the description link says page not available
Is there anyone else finding difficulties while installing punkt in NLTK
Hi Sir, I am getting the follwoing error while importing gensim library:
*ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject*
Please help me to solve this error
where are you getting the code from? i tried the link in the description box its showing 404 error. can you help
Is ineuron link working ?
no
finished watching
Ok