Timeline of Assumption of linear regression 0) 00:51 Introduction 1) 01:58 Linear relationship (between all the independent and dependent features) 2) 04:25 No multicollinearity (between independent features) 3) 11:37 Normality of residuals (Distribution of residuals should be normal) 4) 13:15 Homoscadascity (residuals and predicted values should not have any pattern ) 5) 15:32 No autocorrelation of residuals
Sir aapne bohot hi badhiya padhaya ....after watching a lot of videos but not founded a video like this ....now i understood the concept only by you ...ty
1:44 During train_test_split, the rows of the data are randomly ordered (unless you set a parameter not to reorder, which is not set here). Because of this, the residuals at 16:08 will always show no auto-correlation even if it was, as the order is jumbled up.
Hello Sir, As always this video is also amazing no doubt about that. But in this video you only explained How to check the assumptions. But what if the assumption is not hold than how to tackle them ?? Like what are the processes in order to convert the data so that it holds all assumptions. please make video on that and explain that.. Thankyou Sir
The only video on YT that explains assumptions of alogorithm. Thank you so much sir this video was a great help. Sir, can you please make videos like this for other alogorithms also.
The only missing thing was the "why" Why do we need these assumptions of linear regression. You only explained Multicollinearity would have been perfect if explained for all.
Hi sir, The 1st assumption of linear regression is that the equation should be linear in parameters and there is no restriction on how x and y are related. But u showed in ur video that if there is non linear relationship between x and y then the equation doesn't holds the assumption which I think is not right.
1. Linear relationship between input and output 2. No multi collinearity 3. Normality of residual 4. Homoscedasticity 5. No auto correlation in residual
Sir in Autocorrelation of Residuals if we sort the data then it will also follow some pattern. This plot depends on the order of input and we can pass input in any order. btw great video. Thanks
Conceptually what does auto correlation of residuals represent? You explained nicely why there should not be a correlation b/w Independent variables. But I didn't understand significance of no-auto correlation assumption for residuals
It would have been a good video , if the reasons behind these assumptions is well explained. the reasons behind the Normal Residual, Homoscedasticity, No Autocorrelation of Error are not explained. how does these assumptions impact the model is not explained. Thanks for explaining the meanings of these errors with examples.
Sir i am facing a bimodal residual issue or problem dont know what to say. Even my teacher dont helped me in that. Can you give some points or anything
In both assumptions 3 (normal residual) and 5 (autocorrelation) , we are plotting residuals, How come assumption 3 says it is normally distributed but 5 says there is not relation?
Isn't the assumption about linearity in the sense that it should be linear in parameters, not variable? That is to say the assumption is fine with non linearity in X variables until coefficient (ß) of X is linear.
I would suggest you to explain why linear regression assumes, normal residuals, homoscdacitiy and no correlation between/w residuals/independent variables. That will help make your channel different from others because it will help your audience understand the concept better. The things you have explained, anyone can explain it, but only a handful number of people explains "why"
hello there. As u said these are the assumptions in LR and a candidate who is not aware of these is judged. But the thing is, from where can one read about such concepts?. Can you please suggest some books with solid ml fundamentals as there is a lot of ambiguity about concepts in ML books and not every book talks in depth about these algorithms.
@@campusx-official thank you sir. Sir is macbook air m1 good for such work such as ml and dl I am watching your 100 days of machine learning and reached day 3 because it been only 3 days of me starting this new journey
You should actually rectify yourself. The linear assumption is never about straight line... it is linear in estimation parameters. y= k x^2 is also linear regression. Please correct.
![[TF2] The Russian Bear Trap](http://i.ytimg.com/vi/YFNSwz2_rOA/mqdefault.jpg)
Timeline of
Assumption of linear regression
0) 00:51 Introduction
1) 01:58 Linear relationship (between all the independent and dependent features)
2) 04:25 No multicollinearity (between independent features)
3) 11:37 Normality of residuals (Distribution of residuals should be normal)
4) 13:15 Homoscadascity (residuals and predicted values should not have any pattern )
5) 15:32 No autocorrelation of residuals
Sir aapse accha koi nhi samjha sakta ❤️
I truly appreciate your explanation; it's been incredibly helpful. Thank you very much, sir!
Sir if possible make video on AUC ROC curve...and thank you for making this video
Sir aapne bohot hi badhiya padhaya ....after watching a lot of videos but not founded a video like this ....now i understood the concept only by you ...ty
This Question has been asked in Turing data scientiest interview Sir, thank you so much.
thanku bhai
the most "to the point" and easiest explained video in youtube.
Very Helpful Video for people who grasp a concept from fundamentals. Very intuitive with practical implementation.
Sir,ur teaching skills is very awesome ❤️It was much helpful for me Thanku 💐🎉
This is what content is. I hope you will give deep understanding on other topics too
Yours videos are very well explained ,Thank you soo much Sir for giving the knowledge. You are the best teacher ever
Thank You Sir, the way you make us understand is really great... Love From Pakistan 💖
1:44 During train_test_split, the rows of the data are randomly ordered (unless you set a parameter not to reorder, which is not set here). Because of this, the residuals at 16:08 will always show no auto-correlation even if it was, as the order is jumbled up.
Thank you for this wonderful information🎉
Great explanation sir,simple illustrated by example
Hello Sir, As always this video is also amazing no doubt about that. But in this video you only explained How to check the assumptions. But what if the assumption is not hold than how to tackle them ?? Like what are the processes in order to convert the data so that it holds all assumptions. please make video on that and explain that.. Thankyou Sir
Dude, you are really a genius......excellent explaination
The only video on YT that explains assumptions of alogorithm. Thank you so much sir this video was a great help. Sir, can you please make videos like this for other alogorithms also.
One of the best and quick video I have seen
ON THE POINT!!!!! VERY IMPORTANT INFORMATION REGARDING INTERVIEWS
Amazing Knowledge Sir.
Please make video on how to overcome each assumption if it is invalid
sir videos is very good , sir we need a videos for that case if assumption not satisfy how we can use remedy of these assumption in python
Best teacher ever!
Good Video on Assumptions of Linear Regression🙂
Thank You Sir.
Very nice explanation❤
you simplified the concept..Thank youuuuu
The only missing thing was the "why" Why do we need these assumptions of linear regression. You only explained Multicollinearity would have been perfect if explained for all.
Thank you. Very helpful
Thanx for clear explanation it was quite informative
Hi sir,
The 1st assumption of linear regression is that the equation should be linear in parameters and there is no restriction on how x and y are related. But u showed in ur video that if there is non linear relationship between x and y then the equation doesn't holds the assumption which I think is not right.
Its very very good lecture sir thanks a lot
thank you clearly explained the concepts
Same Question asked me.
Can you explain all assumption of all algorithms??
Plz sir that will be very helpful for us. 🙏🙏🙏🙏
1. Linear relationship between input and output
2. No multi collinearity
3. Normality of residual
4. Homoscedasticity
5. No auto correlation in residual
thanks sir
Thanks you for simple and great explanation
Hi Nitish there is one more assumption for this our response variable should be normally distributed please explain thats why we use GLM
Nice explanation
Well Explained
deep learning video sir!
NIcely Explained
Thankyou Sir.
I will reference this content.
Sir in Autocorrelation of Residuals if we sort the data then it will also follow some pattern. This plot depends on the order of input and we can pass input in any order.
btw great video. Thanks
Very nice explanation sir
Very nice ...
Very well sir, thank you so much 😊
Thanks bro
good one sir
great explanation... thank you! 😇
Nice video sir
Thank You So Much
Thank you sir ♥️
Thank you sir❤❤
thank you so much sir
Great Explanation.. But can you please make a video on detail explanation of autocorrelation and homoscedasticity??.. Thank You
Thank u sir
Make video on
What to do if these assumptions get violated
Thanks jitu bhaiya.
bravo!
Conceptually what does auto correlation of residuals represent? You explained nicely why there should not be a correlation b/w Independent variables. But I didn't understand significance of no-auto correlation assumption for residuals
🎉🎉🎉
It would have been a good video , if the reasons behind these assumptions is well explained. the reasons behind the Normal Residual, Homoscedasticity, No Autocorrelation of Error are not explained. how does these assumptions impact the model is not explained. Thanks for explaining the meanings of these errors with examples.
Sir, aapne bataya tha January me NLP ka playlist khatam hoga ! Abhi June khatam hone wala hai 😞
Excellent content Sir but I have a doubt. Residual should be bell-shaped then how it's not holding any auto-relation correlation?
Awesome 👏
Thankyou sir 🙇
Nice video sir.., the no autocorrealtion assumption is only for linear regression or it is applicable for other algorithms
Thank Men
Sir apne btaya how to check linearity, but ye nahi btaya agar non linear h to krna kya h
Sir please keep making Deep learning videos for the 100 days ml playlist
Sir i am facing a bimodal residual issue or problem dont know what to say. Even my teacher dont helped me in that. Can you give some points or anything
In both assumptions 3 (normal residual) and 5 (autocorrelation) , we are plotting residuals, How come assumption 3 says it is normally distributed but 5 says there is not relation?
thank you sir , apne iski githhub link di hoti to time save hota hamhara...
Isn't the assumption about linearity in the sense that it should be linear in parameters, not variable? That is to say the assumption is fine with non linearity in X variables until coefficient (ß) of X is linear.
LR
No multicollinearity
Normality of residuals
Error should have constant variance
No auto correlation of errors
I would suggest you to explain why linear regression assumes, normal residuals, homoscdacitiy and no correlation between/w residuals/independent variables.
That will help make your channel different from others because it will help your audience understand the concept better.
The things you have explained, anyone can explain it, but only a handful number of people explains "why"
You can explain the reasoning behind it. Others can also chip in. Nitish can correct our understanding if there are any gaps.
sir plz Deep Learning and NLP ki playlist complete kr do.
nice
hello there.
As u said these are the assumptions in LR and a candidate who is not aware of these is judged. But the thing is, from where can one read about such concepts?. Can you please suggest some books with solid ml fundamentals as there is a lot of ambiguity about concepts in ML books and not every book talks in depth about these algorithms.
www.amazon.in/Elements-Statistical-Learning-Prediction-Statistics-ebook/dp/B00475AS2E
What if these assumptions get violated ?
Sir how to handle multicollanirity??? should we drop one column???
yes , we have to drop (highly i think )
but why these assumptions??
SIR how can I join your online 6 month Ml &AI COURSE?PLEASE REPLY SIR.Thank you🙏🏻
Can you make tutorial on deep learning
100 Days of Deep Learning: ruclips.net/p/PLKnIA16_RmvYuZauWaPlRTC54KxSNLtNn
@@campusx-official thank you sir.
Sir is macbook air m1 good for such work such as ml and dl I am watching your 100 days of machine learning and reached day 3 because it been only 3 days of me starting this new journey
You should actually rectify yourself. The linear assumption is never about straight line... it is linear in estimation parameters. y= k x^2 is also linear regression. Please correct.
Wasted 19 minutes . You should explain the reason for having these assumptions
🤡🤡
agar result 1 se 5 ke bitch mai aarahe ho to "multicolinearity hai ya nahi"???????🤨🤨🤨🤨🤨🤨🤨🤨🤨🤨🤨🤨🤨🤨🤨🤨🤨🙄🙄🙄🙄🙄🙄🙄🙄🙄
Thank you sir
Thnks you sir... 👍
amazing sir