Feature Scaling - Standardization | Day 24 | 100 Days of Machine Learning

[@deeptisharma90]

Krish Naik sir is a competent and knowledgeable data scientist, but you are a competent teacher

[@MuhammadYahyaKhan-r2m]

nhi smja

[@ayushmeharkure5365]

englis englis

[@somyaagarwal2942]

Agreed. I like him better than him tbh

[@kashifjalal8616]

Some people are like angels 😇. They just want others to take some light from them and brighten their lives.

[@rohitmeharcse9772]

can anyone help me on this because as per now I am getting very different accuracy which is 87.5 for lr as well

[@neosamosa]

​​@@rohitmeharcse9772accuracy is not dependent only on the code but various other intrinsic stuff. you can try to change and tune your hyperparameters work with the data preprocessing your accuracy can increase.

[@shubhamkumar-hx1fb]

@@rohitmeharcse9772 actually trained and test datas are randomly picked up that's why you are getting different accuracy...i think this could be one of the possible reason for that

[@manujkumarjoshi9342]

Frankly speaking, every data scientist knows it but the best part is you connected all the loosely coupled dots. Best Teacher, I wish I met you before.

[@snehalhon]

You are reallyyyy best teacher!!!!!..... Really after learning from this video no need to watch another video ....you make it crystal clear ...thnxxxx brother ...god blessss youuuu ..

[@learnenglish699]

sir ye sab eda ke bad karnaa hai ya phale he eda kar denaa hai fr data ko clean and then ye sab ????????

[@kunikakhobragade6953]

@@learnenglish699 he answered this in the beginning of the video

[@mahasinprodhan4322]

Didi,Kya aapne yeh machine learning course poora kiya hai?"

[@ShamaAmeer-i4x]

Best teacher for machine learning in whole KZbin channels

[@EN21CSAP]

Let me correct you, not for ML but for Data Science!!!

[@mahasinprodhan4322]

​@@EN21CSAPwhat do you mean Is this playlist for machine learning or data science? Please let me know

[@FahadKinqBtdbBs]

@@mahasinprodhan4322 did you watch his ml engineer vs data scientist video you will understand what this nplaylist is for

[@ajaykushwaha-je6mw]

Guru Ji, charan sparsh!

[@curiousboy7015]

bhai iska english translation dekhna 😂😂

[@fit_tubes_365]

Course Started : ML Lecture-01: 14/08/2024 Lecture-02: 14/08/2024 Lecture-03: 14/08/2024 Lecture-04: 14/08/2024 Lecture-05: 14/08/2024 Lecture-06: 15/08/2024 Lecture-07: 15/08/2024 Lecture-08: 15/08/2024 Lecture-09: 15/08/2024 Lecture-10: 15/08/2024 Lecture-11: 16/08/2024 Lecture-12: 16/08/2024 Lecture-13: 17/08/2024 Lecture-14: 17/08/2024 Lecture-15: 18/08/2024 Lecture-16: 19/08/2024 Lecture-17: 20/08/2024 Lecture-18: 20/08/2024 Lecture-19: 21/08/2024 Lecture-20: 21/08/2024 Lecture-21: 22/08/2024 Lecture-22: 22/08/2024 Lecture-23: 23/08/2024 Lecture-24: 23/08/2024

[@kmh_edits_]

will catch up with you soon.

[@MWASI-kk8nn]

What a teacher in Data science Domain, Very very very nice

[@SumanPokhrel0]

What a beautiful video sir. Clear, to the point, and most of all, the geometrical intuition and code demonstration make everything really easy to understand.

[@AkashKumar-hu7cj]

You are amazing person, You are the motivation factor to learn complex things with ease, I am glad I found this youtube channel for my preperation for machine learning. I feel this channel should reach million subscriber base.

[@festivalvibes9601]

You are amazing person, You are the motivation factor to learn complex things with ease, I am glad I found this youtube channel

[@749srobin]

is explanation mein rishta hone ka potential hai bhai . Bht hi bdiya, Dhanyawaad

[@himadrishekharroy4257]

One of the best content in KZbin till date. Thank you very much

[@tusharbedse9523]

Loved this video and will never open youtube for watching this concept again... you made it crystal clear!!!!!

[@ThaCoders]

amzing class no word to say ur work and dedication hats of u sir

[@MarufKhanPGICS]

you are really the best teacher the way you are teaching absolutely it is awsome

[@abinashnath2481]

Wonderful explanation with example. I am totally clarified with all those concepts after this video.Thank you

[@ritwikgupta7540]

Amazing explanation!! One of the best videos on feature scaling so far...way to go sir!

[@ArunKumar_237]

best teacher I ever seen in ML. teaching to the core

[@surajghogare8931]

U r the complet teacher on KZbin for DS🙏🙏😘

[@usamaabid9190]

Got your channel in recommendations today and really loved your content. Its amazing... Thanks alot♥️

[@mahasinprodhan4322]

Dada,Kya aapne yeh machine learning course poora kiya hai?"

[@MuhammadJunaid-yr8jd]

You are reallyyyy best teacher!!!!!

[@parulpurwar3027]

This playlist is GOLD

[@rohitvishwakarma2871]

For some reason i got better accuracy score with unscaled values than scaled values with Logistic regression. I am using this same example of Age and Salary data

[@RupeshVarma-d1h]

same issue with me also

[@narvarstrange1920]

@@RupeshVarma-d1h i think the algorithm or library has been updated recently... so it results in better performance even without scaling values, so no issue here.

[@talhaanik8441]

same issue: actual accuray 0.875 scaled accuray 0.8666666666666667

[@dhavalahir5785]

You are best for machine learning

[@hamdansiddiqui3294]

Words are come short to explain your channel praise

[@aanchalsharma2567]

Outstanding explaination.. Dhnaywaad sir 🌸

[@prashantmaurya9635]

Hats off to you for this explanation.😍

[@deepanshudutta4443]

I am totally amazed to see your knowledge and teaching style,I really love that.Sir is it possible to learn from you personally ? Or if you have any student batch for machine learning.

[@mahasinprodhan4322]

Dada,Kya aapne yeh machine learning course poora kiya hai?

[@BiswajitDas-lk7pp]

Best Playlist on KZbin

[@iam_shiv07]

Thanks you so much sir, I bet no one can teach ML topics so easily as you do ❤

[@migoswami196]

Really you are great sir. Happy Teachers day in advanced

[@saurabhdas2234]

"Thank you for explaining Standardization so clearly! I finally understand it now.

[@satya3193]

please never stop teaching us, awesome content ❤💘💘💘💘💘❤❤❤❤❤

[@HustlersHubMenu]

🤩🤩🤩🤩🤩 You are doing an amazing work ..............for those who are willing to learn and can learn by themselves/////////////////.............Thanks a lot.......................NITISH SIR

[@pritishpattnaik4674]

Wow sir , crystal clear explanation cleared all my doubts

[@sid_x_18]

Geometric intuition you explain is top level ❤

[@HimanshuSharma-we5li]

You are just soooo good....I wish ...mujhe ye videos pehle mil gaye hote.🙄🙄🙄

[@kunikakhobragade6953]

U r superrrrbbbbb yr🤩 awesome teaching..... everyone should watch ur videos....

[@maitrijain7758]

Really you are a gem for students

[@techvideo4752]

thanks for making such video.

[@tanveerbashir8393]

love you sir🥰 I am enjoying your lectures and learn a lot of things properly

[@because2022]

Great explanation and awesome presentation.

[@abhisheksharma10600]

This series is gold!!

[@monikrayu2546]

no its silver

[@mahasinprodhan4322]

Dada,Kya aapne yeh machine learning course poora kiya hai?"

[@abhisheksharma10600]

@mahasinprodhan4322 haan

[@mahasinprodhan4322]

@@abhisheksharma10600 dada kya yea playlist pura complete hain?kyun ki playlist main 133 video to hain lekin DAY-66 tak ka mention hai thumbnail baki ka video ki video main mention nahi nahi

[@abhisheksharma10600]

@mahasinprodhan4322 each and every video of this playlist is useful and it is related to machine learning so after 66 videos the remaining videos are also related to machine learning and are useful

[@ramuvv9928]

very knowledge person. excellent teaching

[@SaadKhan-gf9ui]

wow man....amazing explanation❣

[@ahmadyaseen2617]

I've taken udemy course In machine learning but his videos are more easy to understand

[@nabajyotidey5613]

u earned my respect^infinity @ Campus X............

[@rutvikkapuriya2336]

best way of teaching thanxxx

[@Scooterboy_and_others109]

Great teacher

[@swethasree5684]

Great Explanation Sir. Thank you

[@narendraparmar1631]

Very helpful Thanks😀

[@asadnawaz6931]

amazing explanation sir

[@zaidnadeem4918]

Best teacher

[@grithijain132]

Best efforts

[@SubirSantra-x3z]

append method is changed to _append, So the correct method of adding new row is df=df._append(df1)

[@dilipgyawali1776]

well explained ...thank u very much

[@zkhan2023]

Awesome video sir

[@adarshvlog2409]

completed it!!

[@percyjackson1662]

I dont know Linear regression or Decision Tree and have just been following this series to learn ML from scratch (coming from DE background) , so my question is - do i need to go and learn all of that first or will they be covered later in this series and i just blindly follow to get atleast basics ?

[@Tech_Charla]

I wish I start learning ML 2 years ago, Still I can catch-up ....

[@ashutoshyadav8653]

its amazing ...

[@mazharuddin9078]

Logistic regression require scaling / as you explained in the example

[@doptopgaming9569]

can anyone tell should i follow this playlist in 2024

[@gamingprincess5461]

yup,its still relevant.

[@SonuK7895]

Bro u are so underrateddddd

[@annyd3406]

what actually np.round(X_train_scaled.describe(),1) at 19:16 did here i couldn't understand it

[@acceleratedofficial1576]

After transforming it gives array as output and we are just seeing it's description

[@jahadhassan3455]

very nice sir

[@balavikaram8547]

Great sir

[@RAHUDAS]

Great , I learnt a lot. Just had one small doubt, you use same instastance of StandardScaler for train and test data, Now my question is that not going to lead *Data Leakage*. ? My intuition is , mean and std of train data would be shared to test data that is not correct. Because according to rule of the thumb training and test set should kept separated.

[@gamingprincess5461]

You're right! test data is being scaled based on the characteristics of the training data. however In many cases, the training and test data come from the same distribution: If the training and test data are drawn from the same distribution, sharing the mean and standard deviation is not a significant issue and the model's performance is not significantly affected.

[@sayansaha5348]

awesome video loved the way you explained .Can you recommend a book to read for machine learning ?

[@campusx-official]

kzbin.info/www/bejne/qXLdiZKje52qfdE

[@sandipansarkar9211]

finished watching

[@arpitchampuriya9535]

Thank you sir

[@SACHINKUMAR-px8kq]

Thankyou so much Sir 24 Day Not Out ...😁

[@usmanriaz6157]

17:10 so you are saying that we use mean and std of X_train on X_test? doesn't make sense.

[@deveshtyagi2996]

how do we get S.D. for xi-x bar at min 8:25

[@swetaacharya2969]

sir plz could you upload the notes you are showing in video ,it will be very helpful 😇

[@monikrayu2546]

luv you sir

[@arun5351]

Great video.. I believe Standardization would be beneficial in all regression based algorithms?

[@campusx-official]

Not all, does not affect regression trees and other tree based algos

[@simranagichani4943]

Please explain why only transform to be used and not fit on the test data. I tried searching it but couldn't understand. I would be grateful. The series is amazing.

[@gamingprincess5461]

If we were to use fit on the test data, we would be learning new parameters from the test data, which would lead to data leakage. Data leakage occurs when the model is exposed to information from the test data during training, which can result in overfitting and poor generalization performance. By using fit only on the training data and transform on both training and test data, we ensure that the model is not exposed to any information from the test data during training, and we get a more accurate estimate of the model's performance.

[@tusarmundhra5560]

awesome

[@learnenglish699]

sir ye sab eda ke bad karnaa hai ya phale he eda kar denaa hai fr data ko clean and then ye sab ????????

[@ajaykushwaha4233]

Hello everyone, I have doubt: Suppose we have encoded categories data with value 0,1, do we need to standardise that features as well ?

[@studyai7832]

Even if you standardise , your output will be 0 and 1 only

[@rohanchitale7294]

Before SVM's also we have to apply Standardization right?

[@arshad1781]

Thanks

[@sliimjiim]

Hey I have a doubt. When I'm getting the accuracy score for normal X_train I'm getting 0.875 and for X_train_scaled I'm getting the score 0.86666667. Could you please tell me what's wrong here? I'm using the same dataset that you used, and already checked all the steps and code thrice! Thanks

[@AlphaRane]

same doubt

[@AlphaRane]

@campusx

[@fit_tubes_365]

re run karo code to aa jayega

[@108_sahilgaikwad7]

hey where can i learn more about x train y train xtest y test? in which video have you explained it sir?

[@karanparashar6824]

But is it appropriate to train a ML model with a Z-scale normalized feature that has negative ages?

[@gamingprincess5461]

Most ML algorithms, like neural networks, decision trees, and random forests, can handle negative values just fine. But, there are some exeptions: Negative ages can make it harder to understand feature importance or coefficients. For example, a negative age coefficient might be confusing. In some cases, negative ages just don't make sense. Like, in life expectancy models, negative ages are nonsensical. Make sure your model can handle domain-specific constraints.

[@Kalimanzaro]

Why there is no loss of information in data due to standardization ?

[@sudhirjangra9524]

Meri Sir ki Logistic wali value Decision Tree me kyu aa rhi hai or DT ki Logistic wali me -- i have checked the code

[@fit_tubes_365]

same mera bhi aa raha tha, fir se run kiya code to aa gaya exactly same

[@sudhirjangra9524]

@@fit_tubes_365 mera to vhi aa rha hai

[@fit_tubes_365]

@@sudhirjangra9524 Accha!🤔

[@rasikakambli7249]

how can we work with 2 different dataset for train and test

[@pranavkolapkar645]

Is standardization same as bringing the values in normal uniform distribution

[@ehababdulrahmanshariff9196]

Im trying to run this now (the logisticregression) but now im getting accuracy without scaling as 0.875 and with scaling it is going down to 0.866 Is anyone else facing a similar situation or can explain this.

[@sliimjiim]

yes i did the coding today and i faced with the same problem, i think maybe they have updated the dataset on kaggle itself afterall its a 3 year old video

[@ehababdulrahmanshariff9196]

​@@sliimjiim I used the data set from the GitHub repo, I am assuming that some updates in sci-kit lean or some other optimisation is at work which we are not aware of.

[@anujdubey7]

I AM GETTING SAME ACTUAL AND SCALED ACCURACY 0.875 WHY?? Did you find anything??

[@mohammadfarazgoriya5929]

Sir i have a doubt...sir why can't we use { lr = LogisticRegression()} for both actual and scaled data..??

[@tanmaygupta8288]

Shouldn't X_test be scaled by it's own mean and standard_deviation?

[@dhirajbarot4339]

Yes, you're correct. In a very simple way: X_train: Standardize using its own mean and standard deviation. X_test: Use the mean and standard deviation learned from X_train to standardize X_test. This ensures that the scaling is consistent and based only on information from the training set. It prevents any information from the testing set leaking into the training process.

[@tanmaygupta8288]

@@dhirajbarot4339 thanks for the clarification

[@sandeepthukral3018]

@@dhirajbarot4339 but after using X_test mean and std are not 0 and 1

[@shovendubey]

Let me tell you one thing. You are better than UDEMY,COURSERA,,DATACAMP ...

[@pawanraje5510]

sir why this happen with my notebook,same code same dataset but the accuracy of non-scaled is better than scaled

[@Wtk_Ncs]

Hii Sir, could we please get this one note file from which you are teaching ?

[@easieml]

Is there any link for class notes?

[@incentivee]

But Mera accuracy_score score same raha hai accuracy_score Actual 0.75, Scaled 0.75

[@shishankrawat2105]

17:10 "seekhte ho aap training pr but transform dono ko krte ho". WHY? Answer is: Splitting the data into training and testing sets before scaling ensures that your model is trained and evaluated in a realistic and unbiased manner, preventing data leakage and enabling fair evaluation of its performance. Credit - ChatGPT. Thank me later... :)