Excellent! Simple and clear explaination

Excellent explanation. Thank you!

wow, I like your teaching approach

Thank you so much. MAE, MSE & RMSE has been a major blocker for me and you've cleared things up.

This is great! thanks a lot

Excellent video so amazing thank you so much! - From a noob bioinformatics grad student

great explanation

7:13, log(100) = 2 and log(130) = 2.11394335231 I think we have to take log of 101 and 131 that's why in formula Log(pi+1)log(ai+1) is available.

You are amazing.

Thanks a lot 😊

thank you so much !

Thank you

Thanks a lot, Nutshell description with example.

Sir All explanation is very nice. Easy to understand.

For rmsle to is the value closer to 0 better?

You are a God. Thank you!

Thank you very much for sharing this video. Indeed you have made it so simple that everyone, in my opinion, will be able to find out these metrics now. But what I don't get is that what do they communicate, i.e. if suppose I found MAE=0.0708 and MSE=8e-5 and RMSE=4e-3, then what can we conclude from these readings about our data?

@akhilendrasingh

4 жыл бұрын

Lower values are better so a model with lower Mae or rmse will be considered better than a model with higher values. It validates your model

@sadiq114

4 жыл бұрын

@@akhilendrasingh Thank you very much R/sir. Suppose I have a desired vector u=[1 2 3 4]. Now here the 1st two elements are say for example amplitudes of two signals and the last two values are the phases of those signals. Suppose I use an algorithm for the estimation of these values.Lets say for example I use Genetic Algorithm and run that algorithm 100 times and GA estimates 100 such vectors as u. i.e. Est1=[0.9999 2.0001 3.0001 3.9991], Est2=[1 2 3 4], Est3=[1.0010 2.0000 3.0010 3.9821], .......Est100=[..........]. Now which error should I use for this? i.e. I want to comment on the performance of the algorithm whether it has performed well or not in estimating u. So which error should I use for this? whether MAE, or MSE or RMSE etc. or all of them? And further what other performance metrics can I use for this?

@akhilendrasingh

4 жыл бұрын

@@sadiq114 Try to use more than 1 loss function.If you have outlier in the data and you want to ignore them, MAE is a better option but if you want to account for them in your loss function, go for MSE/RMSE.

@sadiq114

4 жыл бұрын

Thank you very much R/sir for your kind guidance. But I don't know what do you mean by loss function. Are MAE, MSE, RMSE etc called loss functions?

Thank you Sir.

Thank you so much.

Thank you so much for breaking down the differences. This helped so much.

Excellent video

even RMSE, MSE does not account for negative errors?

IF a Regression Model said to be performing well using performance metrics MAE or MSE, then what will be the ranges of MAE or MSE when data is not scaled? What will be the ranges of MAE and MSE if the data scaled in between 0 and 1 or -1 to 1?

@akhilendrasingh

4 жыл бұрын

Values can range for 0 to infinity where lower value is preferred. Ideal value would be 0 indicating no error but that is practically not possible.

I have a question, should RMSE be less or greater than standard deviation? Or should it equal that of standard deviation ? For example, a dataset with standard deviation 1.915, and after applying linear regression has a RMSE value of 1.909 on the test set and after using Ridge regression it's RMSE is 1.826. Is this considered a good model or not? I Would be grateful for your feedback. Thank you. Regards

nice sir ,,,, did u give any lecture in learning mall?

great

At 1:15 how come the predicted values not on the best fit line? Doesn't it beat the whole purpose?

@akhilendrasingh

3 жыл бұрын

If there is no error, it indicates over fitting. Whole idea in a model evaluation is to identify the errors and reduce them to deliver optimal performance but most models will have some kind of errors.

RMSLE typed wrongly at 6:57 min

mast samjghaya hai

Thank you! Can you please check the link as it is not working

@akhilendrasingh

2 жыл бұрын

thanks, please check now. it is working

how do I know whether my data have outlier or not?

@akhilendrasingh

3 жыл бұрын

Simple explanation for outlier is that they are far away from the normal distribution for example if most values in your dataset are between 50-80 but one or few values are around 150. These values around 150 are your outliers. If you are using r or python, you can print summary of your dataset, that will give information about normal range and outliers

6:40 RMSLE equation miss a right brace some where

@akhilendrasingh

4 жыл бұрын

Thanks for pointing that out, I will check.

MSE CAN NEVER BE NEGATIVE