Top 9 Performance Evaluation Metrics | Machine Learning Classification
There are many evaluation metrics to choose from when training a machine learning model. Choosing the correct metric for your problem, the data at hand, and what you are trying to optimize is critical to the model's success.
This video will teach you about the top 9 machine learning performance evaluation metrics used for classification tasks. Which one is better and when to use what?
If you want to learn about a specific metric, here is a timeline to jump into:
00:00 Introduction
01:16 What's the Point of Evaluation?
02:15 Confusion Matrix
03:24 Positive Vs. Negative in Statistics
04:33 Accuracy
04:59 Is It a Good Measure?
06:20 Recall
06:49 Is It a Good Measure?
07:58 Precision
08:24 Is It a Good Measure?
09:30 Precision VS. Recall
10:06 F1 Score
10:43 Is It a Good Measure?
11:23 Specificity
12:29 Is It a Good Measure?
12:44 Precision VS. Specificity
13:20 Area Under the Curve (AUC)
13:39 Receiver Operating Characteristic Curve (ROC)
15:09 Is It a Good Measure?
16:58 Gini Coefficientnt
17:43 Jaccard Similarity
18:14 Is It a Good Measure?
19:49 Log Loss
20:46 Is It a Good Measure?
21:46 The Best Evaluation Metric?
My original Medium article is here if you would like to read it (it is FREE as thanks for watching this video): https://medium.datadriveninvestor.com...
Keep in mind that regression models need to be evaluated differently because they operate on a different principle compared to classification in machine learning. Let me know if you want me to make another video specifically for the regression models.
Don't forget to smash that like button to help the YouTube algorithm and subscribe to the channel to stay up-to-date with upcoming videos.
Stay safe and Sound
========================================
Ways to connect:
Subscribe!
https://bit.ly/RBrilenkovYT
LinkedIn
/ ruslan-brilenkov
Medium
/ ruslan-brilenkov
========================================
Disclaimer: everything presented in this video is my own opinion and is meant to educate and share information, nothing mentioned or described here is legal or financial advice. Ruslan Brilenkov is not responsible for any profits or losses associated with your investment. So, please be responsible for your own actions.