Hi there. k-Fol d is useful in situations where you want to assess the performance of your model multiple times without touching the test set. Because of its nature, you will test your model say 5 times (when k = 5) and get 5 measurements. With these measurements, you can see if your model is performing consistently or there is a large variation in performance. So along with the normal accuracy measures, you get back measures of variance (or standard deviation of those measurements). A good model will have consistent results across all k folds. This test of consistency is difficult to do with only one test dataset. After k-fold and you are happy with it's consistent performance, you can then test the model on the test set.
@@PrasadDraws they are standing in for your features. Imagine your dataset had features or columns called x and y. Put together (x, y) is a data point.
Easiest demonstration of K-Fold method ever. Thanks
Thank you, this was very helpful
hello, when we do k-fold CV, does it mean that we never touch our actual testing set? Since we treat our validation set as our "testing" set.
Hi there. k-Fol d is useful in situations where you want to assess the performance of your model multiple times without touching the test set. Because of its nature, you will test your model say 5 times (when k = 5) and get 5 measurements. With these measurements, you can see if your model is performing consistently or there is a large variation in performance. So along with the normal accuracy measures, you get back measures of variance (or standard deviation of those measurements). A good model will have consistent results across all k folds. This test of consistency is difficult to do with only one test dataset. After k-fold and you are happy with it's consistent performance, you can then test the model on the test set.
@@IQmates ahh that makes sense, thank you!
Thank you .. it was very helpful.. but if you don't mind me asking, what software are you using for your video presentation?
Hi Gisele. I am glad you found it useful.
For my "paper", I use SmoothDraw and to record the screen, I use Movavi. :)
Ok great .. thank u ♥️
what is (x1, y1) here ?
x and y stands for !?
@@PrasadDraws they are standing in for your features. Imagine your dataset had features or columns called x and y. Put together (x, y) is a data point.
@@IQmates thanks for the reply...now its clear!!!
what is your name ?
Hi Furqhan. My name is Houston :)