K-Nearest Neighbor from scratch - Machine Learning Python

[@sohanaryal]

Just found your channel, this playlist along with others is amazing, thanks man

[@alexchristian220]

hey bro I want scratch code of KNNimputer is this helpful or not!!!

[@teetanrobotics5363]

please order the ML playlistchronologically. Amazing Channel. Please dont stop

[@ItsNotAllRainbows_and_Unicorns]

Quite an extensive implementation. Thx. I wonder how it can be implemented with MNIST. I have seen other vids where kNN is usd with MNIST. I wonder what parts of the program would have to be changed.

[@AladdinPersson]

I don't think you have to change anything to run it on mnist. Which part do you think would be problematic?

[@ItsNotAllRainbows_and_Unicorns]

@@AladdinPersson Okay, I looked at the data again. I see how you have two files, data and targets. You're right there should be not much of a problem.

[@ocamposfaria5106]

thank you for the great tutorial! :)

[@cyngtr]

Hey aladdin, im currently getting this issue and im unsure as to what is wrong. i assume my data but it doesnt make sense as to why it is wrong?
File "program.py", line 149, in predict_labels
k_closest_classes = self.y_train[y_indices[: self.k]].astype(int)
TypeError: only integer scalar arrays can be converted to a scalar index

[@CasualGamer669]

Can you help me to implement cosine similarity to knn ??

[@darsh8695]

If I am trying to run it with MNIST how do I read in the data?

[@AladdinPersson]

I guess perhaps the easiest is loading it with PyTorch torchvision or something like that. Actually not sure, haven't tried loading mnist outside of pytorch/tensorflow frameworks

[@AmitDube15]

you can load data from sklearn
*from sklearn.datasets import load_digits*
*mnist = load_digits()*
*X = **mnist.data*
*y = **mnist.target*

[@raphaelprinz304]

i can't import numpy. what should i do?

[@idiomprog]

You could check if you have numpy installed... pip install numpy