How to evaluate ML models | Evaluation metrics for machine learning

There are many evaluation metrics to choose from when training a machine learning model. Choosing the correct metric for your problem type and what you’re trying to optimize is critical to the success of the model.
In this video, we will learn about the most commonly used evaluation metrics for classification and regression tasks.
Get your Free API token for AssemblyAI here 👇
www.assemblyai.com/?...

[@forprogramming8541]

This video helped me to pass the azure data scientist associate exam. Thanks for the video.

[@ahmadebrahem4611]

Thanks a lot I needed this clarification for my presentation

[@DiwasTimilsina]

Wow, thank you so much for these video. I am a software engineer by trade but increasingly big tech companies have ML system design as one of their interview rounds. Your content was amazingly helpful in preparing for those interviews!

[@AssemblyAI]

Great to know Diwas! Good luck with your interviews!

[@narendrapratapsinghparmar91]

Thanks for this informative lecture.

[@louisawang1308]

super clear explanation! I seldom leave comments, but this video totally amazed me!

[@AssemblyAI]

That's great to hear, thank you!

[@xxelurraxx232]

*INSANELY* helpful. Thank you *so* much!

[@AssemblyAI]

You're very welcome!

[@penelopeharo8591]

Thank you so much, this video helped me understand the metrics in the clearest way possible.

[@AssemblyAI]

Glad it helped!

[@sholay7255]

Thank you for this video, Thank you so much. I have questions to ask based on this model evaluation, my questions go like this sir, " is there a way to use the confusion matrix to know the exact datapoint in our dataset our model got wrongly during the predictive system? Also, Sir when we deploy our model to a web app using streamlit, can we use a confusion matrix to figure out which exact datapoint our model predicted wrongly by applying the confusion matrix to the final predictive system output in the web app ?

[@dr.dwight]

Awesome content, but something unrelated question, What are your camera settings? I especially like your camera setup, could you give info on that? What lens, what aperture, and anything else is needed to replicate the same light/room setup. Thanks 🙂

[@daniapy]

Well explained! thank you soooooo much

[@AssemblyAI]

You are very welcome. :) - Mısra

[@user-ph1of6te8z]

How to evaluate clustering algorithms like K-Means and Fuzzy C-Means.

[@josephvanname3377]

When you showed TP/TP+FN, this should be TP/(TP+FN) since without the parentheses to specify the correct order of operations, TP/TP+FN=1+FN.

[@daniocionini7043]

Thank you for the informative video

[@AssemblyAI]

You're very welcome!

[@Lexxus28]

great!

[@jamesadeke9873]

Please ma, can you share the codes you used to plot the true positive rate vs false positive rate graph, PR curve? It looks so beautify and i can't get the exact, please help. Thanks in advance

[@teaknowsoundlab]

thank you dear👏

[@AssemblyAI]

You're welcome! - Mısra

[@kristianfella-glanville]

Good video thanks

[@AssemblyAI]

You're welcome

[@haticeozbolat0371]

Greatt

[@kristianfella-glanville]

What do you think about repeated random data splitting e.g you split the data 80 percent for train and 20 percent for test on a random basis that preserves the class structure vs k fold cross validation? Edit yep I now know this is worse

[@katerina_lina]

Why??? I am concidering It now.

[@aymanekanaoui9604]

how to calculate the coefficient of determination (R2)

[@amanpatil441]

❤️

[@DaniyarRunning]

Can you please upload code shortcuts of this metrics ? Thank you in advance

[@AssemblyAI]

Depending on the language you're using, you'd need different resources, but for Python and the Scikit-learn library, here is some really good and comprehensive documentation, explaining the implementation of each metric: scikit-learn.org/stable/modules/model_evaluation.html

[@moab2192]

Good explanation + beautiful face (:

[@sultanaaa8282]

"شكرا جزيلا لك" This means "thank you so much" in Arabic 🕴

[@michaltrodler195]

Idk i somehow seem not being capable of following this pace

[@andresolbach]

How do you pronounce "Accuracy"? Its so triggering lmao Thats not american nor british english, right?

[@greenbrothersuk]

Well explained but might be good if you added something on threshold setting for binary classification

[@AssemblyAI]

Thank you!