Fantastic video, this is something a lot of people struggle to learn in school. Clean, well-done explanation, perfect for me, interesting example you used as well, I'm subscribing! :D
OMG thank you very much, I'm doing Bootcamp and they throw like 10 models at you without explain the logic behind each. Your video sums it all up nicely.
This is so cool! I had no idea what machine learning looks like in practice. I thought your video was also well-done. The intro was good - a nice call-forward to what you were going to demonstrate later, followed by a short title sequence - simple yet effective. I enjoyed the parts where you put in the "I'm learning" robot. Pro tip for next time: make that clip tighter - really cut it down even tighter at the beginning and end for more of a punch. Also, try experimenting with cutting it off mid-sentence. For example, "I'm Learn-" right back to coding. Because you use it multiple times, the audience starts to expect it. Keep up the great work!
suppose we have a data in numerical values do we still need the encoding? Bascially I want to use the logisticregression for detection of fake GPS signals
Very nice vid, helped me a lot for a uni assignment. I'd only add some tips in the end for the multinomial case. Anyway thanks a lot, very well explained and easy to understand :)
Is there any reason why doing drop_first = True when doing the one-hot encoding is required? Obviously, I understand that you can deduce what class the entry is without having a binary column for class_1st class, but does it mess up the model if we just leave it?
Because you want to avoid multicollinearity, especially if you are doing a linear regression. For example, if you have 3 classes, you can write them like this: D1=1; D2= 0; and you can deduce D3 when D1 and D2 equals to 0. Cheers!
Hi Rylan, thanks for the lovely video, it's fantastic. I'm quite new to python, and was wondering, instead of predicting a specific record, how should I extract and save the predictions for the testing dataset as csv file?
Hello. I'm trying to do this method for the image datasets. But when I use my image It gives an error "TypeError: 'builtin_function_or_method' object is not subscriptable" I solved it. If u guys have the same problem too just add [ ] outside of data
Fantastic video, this is something a lot of people struggle to learn in school. Clean, well-done explanation, perfect for me, interesting example you used as well, I'm subscribing! :D
Perfect. Simple & short.
OMG thank you very much, I'm doing Bootcamp and they throw like 10 models at you without explain the logic behind each. Your video sums it all up nicely.
This is so cool! I had no idea what machine learning looks like in practice.
I thought your video was also well-done. The intro was good - a nice call-forward to what you were going to demonstrate later, followed by a short title sequence - simple yet effective.
I enjoyed the parts where you put in the "I'm learning" robot.
Pro tip for next time: make that clip tighter - really cut it down even tighter at the beginning and end for more of a punch.
Also, try experimenting with cutting it off mid-sentence. For example, "I'm Learn-" right back to coding.
Because you use it multiple times, the audience starts to expect it.
Keep up the great work!
Underrated Channel
amazing.. thank you so much. I am learning from old tutorial which doesn't show yet the solver for logistic regression.
Nice video! Clean and to the point. THx! :D
suppose we have a data in numerical values do we still need the encoding? Bascially I want to use the logisticregression for detection of fake GPS signals
Very nice vid, helped me a lot for a uni assignment. I'd only add some tips in the end for the multinomial case.
Anyway thanks a lot, very well explained and easy to understand :)
Is there any reason why doing drop_first = True when doing the one-hot encoding is required? Obviously, I understand that you can deduce what class the entry is without having a binary column for class_1st class, but does it mess up the model if we just leave it?
Because you want to avoid multicollinearity, especially if you are doing a linear regression. For example, if you have 3 classes, you can write them like this: D1=1; D2= 0; and you can deduce D3 when D1 and D2 equals to 0. Cheers!
I had a query about why don't we check assumptions in Logistic Regression before implementing it?
Hi Rylan, thanks for the lovely video, it's fantastic. I'm quite new to python, and was wondering, instead of predicting a specific record, how should I extract and save the predictions for the testing dataset as csv file?
@@pythonmaraton I see. Thank you for the prompt reply!! ^_^
Coursera codes took me to a shit hole. Thanks Rylan!!
So nice
Hello. I'm trying to do this method for the image datasets. But when I use my image It gives an error
"TypeError: 'builtin_function_or_method' object is not subscriptable"
I solved it. If u guys have the same problem too just add [ ] outside of data
'compact as a pill' tutorial