Tutorial 44-Balanced vs Imbalanced Dataset and how to handle Imbalanced Dataset

[@shahnawazkhan1636]

Great sir there is no need to join any institute to learn the Data Science just follow the Krish Naik sir playlist.

[@sushantapanda4589]

You are a great tutor, the way you are explaining, great to see you holds to the subject. Awesome

[@mayurkhandeshe4813]

your teaching is very effective sir.... very easy to understand

[@abhijitsarkar5946]

Nice series. Get going. The numbers should be 630, 270 and the accuracy exactly 90%. This is the same as your original imbalance.

[@SandeepSingh-tf7ni]

Simplistic approach for Beginners, would really appreciate if you could do demo with dataset(1000 rows ) of 4-5 features, as well please explain Xgboost. Thanks in advance. Look forward to you response.

[@arjyabasu1311]

Upto what ratio should we consider it as a balanced dataset ??

[@arjyabasu1311]

@Kushal Hu what ratio that is?

[@louerleseigneur4532]

Thanks Krish

[@tusharbhatnagar3146]

Can you make video/tutorial on hyper parameter tuning in classification algorithms!! As it has been coming to many interviews also.

[@venkataraomannem6585]

Well-done sir. Thanks for sharing very good to understand everyone

[@praveensingh1234]

Very nice explain, Thanks a lot.

[@manishshukla125]

Thanks Sir, plz make a video for overfitting and underfitting

[@neelpatel3844]

Very informative, thank you.

[@aayushijain2160]

Sir I have a doubt in this question that how to handle imbalanced data-set either by using right evaluation metrics or by these sampling techniques???Please let me know I'm very much confused...

[@cinemascope8847]

aayushi jain SMOTE can be used where we are increasing the minority data. Safest technique

[@vineetsansi]

XGboost will take care of the weights by itself and we don't need to do any weight adjustment manually ... is that right?? Great videos .. thanks for sharing them .. I am sure you will get big number of followers very soon!! I am also applying XGBoost to the DataScience youtube channels that I am following and your channel seems to be getting heavier and heavier weights ;)

[@dr.bheemsainik4316]

Sir, I have data with binary classification output variables. the ratio of classes is 7.5:2.5. Is this balanced data or unbalanced data?

[@gopalakrishna9510]

i really happy with imbalanced and balanced dataset explaination.......

[@sashpatra88]

Krish : Can you put this in MACHINE LEARNING playlist? If I am not missing anything

[@azmathalisyed9114]

Great information,, good explanation.. 👌👌

[@shashankvashishtha9149]

can u please explain the that 2 algorithms xgboost and adaboost?

[@mandarpawar27]

Hi krish Plz upload vedios on interview questions

[@udanial]

From which playlist this video is?

[@DatAcuity]

I am just asking, So, if we use xgboost algorithm for classification problem we no need to bother about class imbalance. Am I right sir.

[@MsRAJDIP]

Can u show the techniques of handling missing data excluding mean,median mode technique. I read that u can use regression or classification technique to find missing values but never seen implemented.

[@prasanthkumar7328]

while doing downsampling as mentioned we will be reducing the points to 100 so which points to be reduced or simply picking in random is also not a good practice so how could we select those 100 points.

[@NinjaAnkit]

what i feel when see you videos explanation, your explanation like as when you communicating in english in your video its feels like you are communicating in regional language ..........thats why i understand more.........you explain most of the difficult terms in simple way. I love your explanation.

[@NinjaAnkit]

and also i remembered most of the concept in long time .

[@NinjaAnkit]

and i also like your videos which i watched.

[@surendranathify82]

very useful . thanks. could you pls post video on PCA and LDA and about Regularization as well. thanks

[@krishnaik06]

Thanks, please check my playlist, video is already there for PCA

[@umang8895]

great video, easy to understand.

[@harshays2873]

sir suppose if i have less data to train my model at the time what i have to do?

[@biswanandanpattanayak1938]

how to handle missing data if data is 1tb or more? please explain

[@satyaranjanbehera5492]

good explanation..Thanks..

[@HarpreetKaur-mn4we]

Very helpful video

[@kakarlanagajyothi4089]

any videos for continuing these

[@joyeetamallik5063]

Can you share python code to implement these up sampling techniques. Is this concept is also applicable to NLP datasets?

[@PhilippHusiA]

If working with tf.keras, add following code to model: 1) from skearn.utils import class_weight 2) class_weights = class_weight.compute_class_weight('balanced',np.unique(y_train),y_train) 3) history = tf.keras.model.fit(x_train, y_train, batch_size=x, class_weight=class_weights)

[@gopalakrishna9510]

can you explian with python codes ?

[@victorxu9634]

great content. would be nice if it goes deeper

[@rahulmahajan6391]

Can we do down sampling in credit card fraud detection dataset?

[@kakarlanagajyothi4089]

Small doubt missclassfication and imbalanced are same or different..

[@kakarlanagajyothi4089]

Pls answer me

[@galisurendra6397]

reply this question please

[@prithviraj25]

Thankyou Sir

[@venkataraomannem6585]

Sir can you please do this same as practically. Thank you sir

[@pruthvigirijala8146]

Discuss everything in upcoming video..? :p

[@niketanjha]

Really helpful 🙏

[@Shylajakarthick]

Thank you so much

[@karndeepsingh]

how to deal with imbalance dataset when we have multiclass in target variable?

[@MasterofPlay7]

use other metrics such as F1 score instead of accuracy...

[@sushedbubai]

Waiting for more interview questions

[@krishnaik06]

You can find the complete playlist on the below youtube url kzbin.info/aero/PLZoTAELRMXVPkl7oRvzyNnyj1HS4wt2K- I will be updating this with all the questions

[@davinderc]

Consider using better whiteboard software. Your written words and numbers are nearly impossible to read in Paint.

[@krishnaik06]

Hi Davinder, feedback taken

[@azad8upt]

It should be 250+ in test not 350+

[@Beyond90Days]

how is accuracy 350/30 ?

[@Beyond90Days]

got it :) 350/380

[@pruthvigirijala8146]

You know..? :p

[@moulidinavahi1498]

How we can downsample data points ?