Training Data Vs Test Data Vs Validation Data| Krish Naik

[@netviz8673]

when data set is given to create a model then train dataset is on which the model is trained and it usually uses 70-80% of the data set. Test data uses 20% of the data set and is used for model performance analysis, prediction is made and performance metrics is checked. Now validation data is used for model hyperparameter tuning. We take the data slightly from the training data set. Using cross validation the validation data is obtained from training data set. CV=5 means 5 experiments. If total training data is 8000 then 8000/5 = 1600 means for every cross validation the starting 1600 data points will be used as validation data and the rest of 6400 data points will be used as train data and the accuracy is obtained for the first experiment. In second experiment when cross validation is 2 then then next 1600 data points will be used as validation data and first 1600data points + next 4800 data points = 6400 data points will be used for train data set and so on. So with every exp the validation data and train data set vary a little.

[@zaafirc369]

Thank you Clearly explained as always ! Is it possible to implement an example of cross validation and hyper parameter tuning on python pls? 🙏

[@sukruthraghavendra7305]

Perfectly explained sir understood very well

[@MuhammadJunaid-yr8jd]

Perfectly explained sir understood very well Thanks a lot sir

[@anjalipes]

Thank you for the lovely explanation sir. I have a doubt in the validation data, once you have taken the mean or Avg of the CV, what do we do next with that mean? how does it help in the performance of your model? Does this impact testing data?

[@murthy33]

Please upload English videos. It is very helpful

[@hunterworrier6769]

Krish bro... U should provide simple code of this all video too

[@richadhiman3300]

Beautifully explained sir...

[@iamalien8327]

sir thanks for creating the videos sir kindly guide which book to refer for making the notes

[@mahantachauhan4461]

Thanks for the such an informative video 👏👏

[@muazamazam7115]

Thank you for making this piece of sweet pie ❤️🌹

[@AlAmin-xy5ff]

excellent

[@neal_indian555]

Sir you are great Akshay Kumar is not Boss You are the BOSS

[@theunscripted643]

Ur so amazing sir ❤

[@123vijit]

What is hyperparameter?

[@jyotikapatil2354]

This was cross-validation. How hyperparameter tuning is done?

[@rohitjana7461]

Thank you sir ☺

[@bivasbisht1244]

Why hyperparameter tuning is done after training the model ? Shouldn’t it be done before traning the model , and then train the model based on those parameters ?

[@fran9426]

You do both. Once you have done hyperparameter tuning you can then train the final model with the tunes values on all the training data

[@jaiprakashsingh756]

sir, could you please share some notes so that we can revice all the bulletin points.

[@divyadewangan6123]

you are the best person to explain any topic .Thanx🥰

[@omdodmani3205]

if test data is used for accuracy , then which accuracy is shown by the cross validation method ?

[@swagatsanketpriyadarsan6231]

I just gave the 1000th like. :P

[@hunterworrier6769]

But very helpful

[@vaibhvkushwah]

kafi kya bahut jyda acha tha

[@abhinavrajsaxena789]

6400

[@AdityaSharma-yc8yt]

Didn't liked it !

[@vibhu613]

[@sketchytv1321]

خوش رہو بہت اعلیٰ کام کر رہے ہیں ۔

[@zaafirc369]

ROC AUC explanation also pls

[@riya-tb6sj]

Hello sir i m 12 Ai frst tym i watched ur video and frst tym understood concept plz make more videos of 12 syllabus cbse

[@milansood8240]

thank you sir, I'm enrolled in upgrad Data Science program and not satisfied with it as the content is not much special like yours.

[@musicallystatus5368]

@9:52 why you divide for cross velidation= 5

[@akshaymondal1372]

Sir also make a video on data warehouse data lakes

[@vasimshaikh9857]

Thank you sir 💖💖💖💖💖

[@maths_impact]

Outstanding sir

[@neal_indian555]

Hukum ka ekka BOSS

[@PS-rh8vy]

Superb ❤

[@rds9815]

Hi krish Please let me know what is x_train ,y_train and x_test ,y_test and how this process works on the data set including the mathematical equation I have a lot of confusion on this. Please make a separate video on this Both languages of English and Hindi

[@sumitlasiwa7152]

train and test data is further divided into independent and dependent features. x-train is for independent and y-train for dependent of train dataset. This is done to let model know the difference between independent and dependent features. Similarly, same thing for test dataset. In test dataset this is done to compare the model prediction and actual value(y-test) to know the accuracy.

[@rds9815]

@@sumitlasiwa7152 Thank you doubt cleared

[@TheNeu2ron]