Correlation and covariance are so similar.
But there is one really important difference.
Let me explain it:
But there is one really important difference.
Let me explain it:
Covariance
The scale of the covariance depends on the scale of the data. It can take any value. No restrictions.
It's hard to compare different datasets with covariance for the reason above.
It's mainly used in Finance and Economics.
The scale of the covariance depends on the scale of the data. It can take any value. No restrictions.
It's hard to compare different datasets with covariance for the reason above.
It's mainly used in Finance and Economics.
Correlation
It is a standardized measure of covariance. The values range from -1 to 1 every time, no matter the scale of data.
This makes it a powerful tool since we can compare correlations for different datasets.
It's easier to understand, so it's used everywhere.
It is a standardized measure of covariance. The values range from -1 to 1 every time, no matter the scale of data.
This makes it a powerful tool since we can compare correlations for different datasets.
It's easier to understand, so it's used everywhere.
Look at this example:
The covariance between weight and height (in meters) is 0.157, and the correlation is 0.706 (strong positive correlation).
What if we convert m to cm?
Covariance becomes 15.7 - changed with scale.
But the correlation remains the same, 0.706.
The covariance between weight and height (in meters) is 0.157, and the correlation is 0.706 (strong positive correlation).
What if we convert m to cm?
Covariance becomes 15.7 - changed with scale.
But the correlation remains the same, 0.706.
To conclude
Both of them are used to measure the relationship between variables.
The only difference is their scale. While the scale of covariance is not uniform, the correlation is easily comperable.
Both of them are used to measure the relationship between variables.
The only difference is their scale. While the scale of covariance is not uniform, the correlation is easily comperable.
Did you like this post?
Hit that follow button for me and pay back with your support.
It literally takes 1 second for you but makes me 10x happier.
Thanks 😉
Hit that follow button for me and pay back with your support.
It literally takes 1 second for you but makes me 10x happier.
Thanks 😉
https://twitter.com/2712332320/status/1764958045443174782
If you haven't already, join our newsletter DSBoost.
We share:
• Podcast notes
• Learning resources
• Interesting collections of content
dsboost.dev
We share:
• Podcast notes
• Learning resources
• Interesting collections of content
dsboost.dev
• • •
Missing some Tweet in this thread? You can try to
force a refresh
