What is Cross Validation and its types?

[@hanman5195]

i did not find proper online stuff on Data science till you come . Now you came i am amazed at watching and learning data science stuff from you. You are my real Guru ,awesome Sir.

[@jammy2003]

Hrithik Roshan who? You're the real Krish! Thanks so much man, saved my ass as I'm reviewing and revising for finals now. God bless and you've earned yourself a subscriber.

[@sshiv908]

just when I started thinking engineering is so difficult, I found you. thanks for such good explaination

[@samixwasup8561]

this is very close to my university course here in Canada, in fact, you have approached it in a simpler way, which is great

[@aadishjain6838]

easily the best data science channel out there! thank you

[@MsRAJDIP]

So Detailed explanation as always

[@ameerhussain5405]

First time watching your tutorial Quality content, clear and crisp Subscribed👍 Keep making more!!

[@harikrishna-harrypth]

Krish Naik, you're a G!!!! Thanks for taking your valuable time to make these videos for us!

[@hejarshahabi114]

Krish you're amazing, to be honest, I have learnt many things regarding data science from your channel keep going, you are doing awesome.

[@waywaller2630]

Krish! This video simplifies the concept cross validation brilliantly. Good stuff Homie!

[@ujjwalsingh888]

Sessional tomorrow. Couldn't be explained better. Thank you, sir.

[@ruzanb433]

You are Exceptionally good, the clarity about the concept is on point ! Dam

[@sachinahankari]

Simple and lucid English with examples makes it easy to understand.. don't worry about comments, be cool as you seems to be and do your job of giving knowledge..

[@rashmimbendakaluru7225]

superb sir, hats off for your simplified explanation with comparison too. very helpful sir, looking forward for videos on other validation methods, thank you once again

[@yuvachandra330]

Crisp and Clear. thank you Krish.

[@huangkevin5183]

Thank you Krish. Have been watching your videos for my uni project. Really helpful contents and clear explanations!

[@prathameshkoli1355]

Thank you so much Krish Sir, As usual Great Explanation😊

[@cha5848]

I have my deep learning exam in few days ur lectures are really helpful for revision

[@sebastianno8]

Great job, finally I understand what is the purpose of cross validation. I appreciate especially the part when you explained how to communicate cross validation to the stake holders. Thanks again

[@saikrishnapadarthi987]

I've been watching your videos on ML. These are simply brilliant. Your explanation is awesome. Could you please provide cards at the right top when you are referring to your other videos. it helps us more. Thanks for your efforts brother...

[@rishabhnegi9530]

thanks it really removes my headache of understanding that split stuff of k-fold

[@varunmajithia2470]

Superb!!!!!!!!!! . Simply Excellent And Very Well Explained.

[@bhushanchaudhari378]

Excellent videos sir.. thanks alot

[@surajkhanna1129]

Binge watching your awesome machine learning playlist!

[@gangaholi8600]

Very nice presentation and very much informative.

[@bhushandeotale3625]

Best explanation got cleared at first instance

[@jasminejubbal5732]

Great videos Krish thank you so much for sharing

[@hassanabdullahi2203]

How comes i didnt subscribe for krish,,Oh man,You got my respect.Thumps up to you

[@chetanapatil8808]

Best Explaination ever!! Thank you Sir

[@aravindvasudev7921]

This video was very helpful. Thank You!

[@radhakrishna.kalluri7192]

Good job Mr. Krishna

[@claudiusandika5366]

Much respect to you, Sir.. for sharing this detailed knowledge for free. May God Bless you always

[@MJXed]

700k subscribers with a board, sir, you are a legend! ❤

[@anandacharya9919]

You are doing great service to students, hats off to you. Can you please mention the order of DS project. Thank you

[@seriouscoder1727]

Krish you r awsome teacher

[@Ali124hdkflc]

Great video.

[@Moneylisey]

very well explained, thank you very much!

[@ashwaniagarwal6828]

Crystal clear explanation !!!

[@osamaakhtar6525]

subscribed! keep up the good work

[@sandipansarkar9211]

Thanks Krish .Great Explanation

[@sonnix31]

You are the best. Thank you.

[@blablawtf]

You explained it SO well, thank you so much!

[@ashinkajay]

Thank you so much.. God bless you too !

[@prodyutdas1474]

Thank you so much for this nice explanation.

[@mahnabraja1086]

very nice informative videos

[@rambaldotra2221]

Amazing Thanks a lot Sir

[@baharylmaz9366]

This video is amazing thanx

[@kadhirn4792]

Thanks a ton. You will reap good.

[@rutdvajrawal7933]

Thanks a lot krish. Do you have a community or a telegram channel

[@shubhangiwaghmare7947]

very well explained.. thank you

[@samithmadusanka7573]

Nice explanation. Thank you so much. Good luck

[@ashishtiwari1912]

Excellent video.

[@kurtissac]

Thank you very much Krish

[@ayushrajput9988]

Thank you soo much sir for this video clears my concepts

[@MK-yl7jl]

Awesome Explanation Man. Thank you

[@MysticAngel3224]

Thank you so much for a clear explanation of this concept

[@mukul98m]

please do make a video on different types of encoding teachniqe

[@rafibasha1840]

Hi Krish,one data means one datapoint or one row

[@rohitmalhotra2383]

Hi Krish..your videos are awesome..the kind of passion and enthusiasm you have for data science and also for empowering others with same knowledge is exemplary. I am trying to learn data science/ML to employ them in Industrial Systems. Pls help me what approach should i follow. Share sources where i can learn to employ them in Industrial systems.

[@MuhammadUmair-or3or]

no github link?

[@thatguyadarsh]

God Bless You Too.. Thanks!

[@aditisrivastava7079]

Amazing explaination..........Thanks

[@chiranjeevibelagur2275]

Hi Krish, nice explanation on CV. But can you show in the codes where i find this accuracy for any kind of CV method. Because I'm a little confused about that part. Or please suggest the video if you have already done that

[@Funtoosh5395]

Please teach multivariate timeseries data for classification and clustering

[@TJ-wo1xt]

gr8 video

[@060584saurav]

Good Explanation however I have one doubt. The model is trained k times with different datasets so, the hyperparameters of the each model will be different. How to determine and see the best model and its associated hyperparameter. Thank You

[@madhurabhalke2344]

So like for every dataset.. Cross validation is mustt???

[@subramaniannk3364]

Sir, the purpose of cross-validation is to reduce the variance, not bias, correct?

[@RinkiSingh-ph6oo]

excellent explanation could u explain genialized linear modal pls👍

[@preetham4974]

Awsome tutorial,Its been a while since i started to follow your videos and the results are pretty much good and easy to understand from base.I have a query regarding this video,At 14:00 what do you mean by both class ? and what do you mean when you said instances of classes thank you!

[@viveksingh881]

so at 14:00 sir was explaining cross validation with example of binary classification that means we have two class so our output will be either 1 or 0....this is what he meant by two class and he meant the same when he said instances of class.... hope u get my point.

[@williamwambua7710]

Hello sir, You inspire me a lot in the data science and ML journey. Plz can i kindly correct something that I have noted in the vid. I believe we cannot calculate the number of K as you have done it bcoz we mostly just choose a suitable number. If am wrong plz bring me back on track. I believe what you were trying to check is the portion size of each partition in a fold, that can be calculated using the size of the train_test split which i bliv in the example you used is 80%20% therefore using the 1000 records ->1000 of 20% =200 in each partition of a fold then the K we choose it to our preference. Plz bring me back on track if am wrong.

[@ghostsleep2852]

Will it be a good idea to suffle the data and then perform k fold CV?

[@21471adnan]

Really nice video..How to decide which CV to apply ( K told or stratified ) ?

[@ankushjamthikar239]

Hi Krish, the video is very intuitive and clearly explained! Just wanted to know, do we need to perform cross-validation during hyperparamter tunning? Or should we do it two times, that is (i) duing hyper-paramter tuning and (ii) then after training the model with optimized paramters? In K-fold cross-validation, we have k models. Which one should we deploy? Thanks in Advanced!

[@JKGMo1995]

you can use Nested cross validation and do both at the same time

[@annperera6352]

Thank you so much sir.

[@minhadtuashimahmed3873]

Have a little confusion is the test and validation set are the same or different

[@sahilsahil-qz2oo]

Great job Sir, Thank you very much ♣

[@simranuppal7171]

Hi, thank you for the video. I was thinking that the fourth method is very close to moving averages. Could we also have done a 5 day moving average for prediction or as an alternate?

[@sivakumarprasadchebiyyam9444]

Hi Krishna, nice video. Could you please let me know cross validation is applied on train data or total data?

[@kotesh634]

the cross validation is applied in the Train data (ex: 80 :20, total observations: 1000, train =800, test=200 , if k=4 then it is 200 each fold, so one of the fold is considered as Cross validation , but each time it will change the FOLD-1,FOLD-2,FOLD-3,FOLD-4 . total 4 times it will calculate the accuracy score then average it and give the final score.

[@tearistovic]

Thank youuu !!!! ❤

[@shafa-atali7816]

thanks a lot for ur video. Can u kindly tell how to find accuracy of a CNN model using k fold cross validation.

[@nitishreddy4888]

Ok

[@bit-colombo5595]

Sir your videos are awesome, can you also do a video on dummification and prediction window size?

[@saivikas569]

detailed explantion about CV

[@manoranjanmartha24]

Teaching quality is good but voice is not audible plz use 🎙️

[@tanishsharma5440]

Thankyou so much

[@lemonade7134]

Low Bias is actually a good thing, right? Because the meaning of low bias is the difference between the actual value and model-predicted value is low. So that's how a model is supposed to be right? (Correct me if I am wrong)

[@SararithMaoplus]

Hello, do you have code implementation for Time series cross validation? Thank you so much.

[@benoitmialet9842]

Thank you.

[@shahbazkhanlabengineermech3246]

Greeting Krish thanks for a wonderful video just wanted to know after doing cross validation and find the CV accuracy do we need to test the model with the test data set or this CV accuracy is the final accuracy of the model and i dont need to the test the model on the test data. Thanks again.. Hope to hear from you soon Stay Safe

[@anirudhashukla1305]

Can some share the link of the project where he has used timeseries cross validation

[@yogeshwarshendye4857]

I am including some a fraction of my training data for validation. Is it a acceptable thing to do?

[@abhinavprasad5743]

hello sir!! Can you please upload the practical example on above mentioned different types of cross-validation...specially on stratified cv

[@trend_dindia6714]

When we apply hyper parameters tuning do it apply on k fold or stratified cv

[@lavanyasagunthala3801]

Its good sir. But why can't we use random sampling without replacement here.

[@ig2947]

Brilliant..!

[@onkarchothe6897]

can you suggest to me the book for cross-validation and k fold validation? which contains with example with solutions?

[@fidaullah3775]

Please tell me what is 10 fold cross validation.? I am confused

[@__looser__5204]

Sir why cant we shuffle the data set and apply the k-fold...

[@harrywadhwa1395]

I have a doubt. For Leave one out cv technique how come we attain the low bias condition ? If the model is in production and new test data is introduced it would yield lesser accuracy which means model is over fitting that is because of high variance in test and low bias in train. So shouldn't this be high variance since we're taking the generalisation error into picture.

[@mr_nb]

Is it test set or validation set, i am confused ... 😕🤔🤔🤔

[@hemantdas9546]

In leave one out what is the problem if it has low bias. Ideally algo should have low bias low variance.

[@viveksingh881]

computational power...it will take a lot of time to process completely....