Bias Variance Simplified |Machine Learning Fundamentals: Bias and Variance | Bias Variance Trade off

[@AshwiniPeacefulEscapes]

Finally...You are the only person who made me understand bias and variance all thanks to you..

[@shikhar_anand]

Thank you very much Aman for that last point. I actually forgot everything about bias-variance in my interview but not anymore with that last point of explaination. Thanks for keeping it simple..

[@UnfoldDataScience]

Thanks Shikhar.

[@syedfayeqjeelani54]

Aman, thank you man for these small and eloquent topics. You are a savior. Stay blessed.

[@UnfoldDataScience]

🍻, thank you Syed.

[@Sagar_Tachtode_777]

Conclusion:- too much Love for training Data ---} High variance model (decision tree etc) 👌

[@UnfoldDataScience]

:)

[@mohdsajidkhan9713]

This is right break through to understand this concept. I will go through video with n number of time. Until or unless. everything fits in my mind.

[@UnfoldDataScience]

Thanks Sajid.

[@ArvindKumar-ml6rb]

Sir.. Ur method of teaching is awesome..

[@UnfoldDataScience]

Thank you Arvind. Keep watching :)

[@oompster]

Thank you very much for your video. helping me a lot for this topic

[@Sagar_Tachtode_777]

So wonderful explanation. Thanks a lot 🌺🌼🌺

[@UnfoldDataScience]

Welcome Sagar.

[@tusharbedse9523]

good video

[@UnfoldDataScience]

Thanks Tushar.

[@user-cb7it9cv9l]

simple and clear explanation .Thank you for the video

[@UnfoldDataScience]

Glad it was helpful!

[@arshveersingh6499]

Thanks Aman Sir, its very nice explanation given by you in this small video .Your video contain all points which are necessary. Making notes with help of your video became very easy for me. That last point about Bias Variance Trade off is awesome 👏.

[@YogeshKumar-ck2ni]

are sir aap ne to kamal kr dia sb kuch samajm m aagya ek baar m thanks you so much for the great lecture

[@UnfoldDataScience]

Welcome Yogesh

[@sandipansarkar9211]

finished watching

[@KrishnaKumar-ni8fm]

Very useful and nicely explained.

[@UnfoldDataScience]

Thanks Krishna :)

[@carryminatifan9928]

This one is very good explanation. I totally understand the concept ♥️

[@UnfoldDataScience]

Thanks a lot.

[@storiesshubham4145]

Unbelievably superb explanation!

[@UnfoldDataScience]

Glad you liked it!

[@ashishmehra5143]

BIAS NOTES:
When the form of the function is biased.
What is the function:
For example, the linear regression equation is y = mx+c. In other words, we can say the mx + c is a f(x).
i.e f(x) = mx + c
What is the form: If we plot the equation, it wil be a line. In this case, the line is the form. (Side note: Form will depend on the function. It could be parabola, cube and so on, it will depend on what function are we plotting.)
Bias Summary: Coming back to the "bias" using the linear regression function f(x), here the underlying assumption is that the "form" will be a straight line, that is what is called the "bias" of the function.
In the other words, since line (form) drawn from the function f(x) = mx + c is biased, we can say that model build using the same function is "High Biased Model".

[@UnfoldDataScience]

Thanks for writing in. Its very important to see the efforts coming from students. Cheers

[@Aniqahns]

Thank you so much for the last point.

[@pawanchadhokar8134]

Thanks sir for sharing knowledgeable information

[@UnfoldDataScience]

So nice of you Pawan

[@bharathy516]

Well explained

[@UnfoldDataScience]

Thank you

[@VinodRS01]

Simply awesome thank you so much sir.

[@UnfoldDataScience]

Welcome Vinod.

[@VinodRS01]

@@UnfoldDataScience sir please make video related to central limit theorem ,

[@sheruloves9190]

I have high variance towards unfold data science

[@UnfoldDataScience]

:)

[@surendrareddy7190]

Way of explanation is superb , everyone can understand easily , excellent work sir🙏

[@UnfoldDataScience]

Thanks Surendra. :)

[@bangarrajumuppidu8354]

superb explanation sir!!

[@UnfoldDataScience]

Thanks a lot.

[@Vinit_Gambhir]

You are a life saver

[@meenakshirao4990]

Thank you for nice explanation in simple manner

[@UnfoldDataScience]

You are welcome

[@kousikdutta5889]

I have been watching your videos sir , your explanations are wonderful.

[@UnfoldDataScience]

Thanks Kousik

[@RamanKumar-ss2ro]

Good one again.

[@UnfoldDataScience]

Thanks a lot :)

[@charanramtejkodi9220]

Well explained sir

[@UnfoldDataScience]

Thanks Charan.

[@tusharbedse9523]

i like the part "love for training data' :)

[@UnfoldDataScience]

Thanks Again Tushar.

[@firefly2008god]

Very good explanation. Please make a video on Ridge and Lasso regularisation.

[@UnfoldDataScience]

Thanks Rahul, video is already there. Please search for "topic name + channel name" I'm UA-cam.

[@firstkaransingh]

Excellent explanation 👌

[@carryminatifan9928]

Sir how we will know that my model is biased or varianced ?
Any practical video is there on your channel or any plan to create one?

[@UnfoldDataScience]

You will get to know on basic of two things first is, how that model fits internally to data and second how the model is performing on unseen data. For ex - Test data.

[@surendrareddy7190]

When your model gives high accuracy on train data and low accuracy on new data(test data) that means our model has high variance .
Both the time it gives low accuracy means our model has high bias .
I'm I right sir.🙏

[@carryminatifan9928]

Thankyou for the reply guys 👍
Now getting clear.

[@secureblob4874]

I am new to data science and trying to learn fundamentals. Your explanation is simply superb. I was trying to find out why we use the term bias and finally your video gave me the answer! As per the video, if there is an under lying assumption about the form of the function, it is called bias. So, are you saying that the assumption was a human error? ie, the machine learning engineer made an assumption that he can use liner regression to fit the data? or could it be a problem with the data itself? for example, the sample of data the ML engineer got appears like it can be fitted with a liner regression and then the engineer used Liner regression. But actually the sample data cud be wrong and it gave a wrong impression that it cud be fitted with a linear regression model, but actually the data set about the population is not really a linear regression. So as the sample was wrong, the engineer chose linear regression model and he saw the error when the model is tested against the test data. Can you explain? By the way i subscribed to your channel and i believe your channel would need more subscribers. Keep up the good work.

[@UnfoldDataScience]

Thanks for asking.
The answer is, nature of data depends on type of business we are talking about. Data may totally be different for a tourism company as compared to Reatil.
Normally, we do not know about data without digging deeper in it however regression models are good to to start with as it is an old and trusted technique.

[@dany28g61]

Clean and crisp superb.. Would you give trainings as well on DS ML DL?

[@UnfoldDataScience]

Thanks a lot.

[@lavanyagowda8186]

Please update interview questions on data science

[@abhinavkale4632]

The dislikes are from all the edu-tech companies

[@UnfoldDataScience]

😃😃kya kare, unko free education acha nahi lagta.

[@virozz1024]

Y = ax + b , b = bias

[@UnfoldDataScience]

Cheers for comment.