Finally a simple, clear and complete explanation of this word embeddings. Thank you so much. I've studied and searched for months and you've just made the whole thing clear and straight forward.
May Allah increase your knowledge my friend, i am just starting with NLP and you help me learn the basics well. I didnt know what one hot embedding means. Thanks and keep it up. Also could you please do a project end to end in NLP for example sentiment analysis ? I want to see these techniques in real action. Thanks again :)
I really don't comment but this is just great. Really helped me out with my college assignment. I was tearing my hair out before this. Genuinely thank you.
I find ur videos very helpful, many topics which are not explained well in other channels, i come to ur channel to see if it is available and u have never disappointed me. Keep up the good work. Thanks.
Precise explanation. Very useful. Thank you Aman. Just one suggestion, visually it will be more soothing if you could follow straight line on your whiteboard.
@Unfold Data Science In the starting whatever encoding ur doing, I don't think it is one hot encoding instead it's a BOW representation...Correct me if i m wrong.
Hi, I have an experience of 8 years in IT(Programming) and started preparing for data science career transition. My query is How can I put up in Resume. Shall I tell them that, I am new to this and Know all the concept and worked on kaggle datasets? Will they allow or reject since no past experience on data science? looking forward for your valuable comments.
Hi, The query needs some bit of understanding your profile and giving suggestion I see you are looking for some career guidance, If yes, you can reach me directly for one on one mentorship. You can go to "About" section in the channel and click the link "One to One Mentorship Link"
Great explanation and well done, my question is how the numbers can be selected to set the features? I mean, in your example at time (5:00) of your video, you assigned 0.9 to the apple and 0.85 to mango ..etc, how these numbers can be assigned?
Sir, I have one query that as input we need to feed the one hot representation of the surrounding words including the one hot vector of the target word or excluding so?
@@UnfoldDataScience okay sir..then in case of getting all the word embeddings of the input sentence..one hot encode of the words under the window passed through CBOW and find the middle target word using the feature based information.. then slide the window and repeat the same untill the end... is it the process sir?
Normally we take 300 dimension however if our vocab size is huge, then processing might be difficult with more number of dimension hence we try limiting it to 200 or so. If you do not have infra issue, you can go for 500 or even more.
@@UnfoldDataScience so it is considered as a hyperparameter?...can we change this in tunning. and please make video on optuna, biasen hyperparameter tunnign.
