CORRELATION
COVARIANCE
REGRESSION.
What are they??
And is their any form of relationship between these 3?
Well let’s find out in 3 minutes ☺️
Retweet …… cos it’s a THREAD 🧵
COVARIANCE
REGRESSION.
What are they??
And is their any form of relationship between these 3?
Well let’s find out in 3 minutes ☺️
Retweet …… cos it’s a THREAD 🧵
CORRELATION
Correlation is used to test for the strength and direction of association between 2 variables
If the association causes the variables to change in the SAME direction, we have a POSITIVE CORRELATION.
If the change is in OPPOSITE direction, we have NEGATIVE correlation
Correlation is used to test for the strength and direction of association between 2 variables
If the association causes the variables to change in the SAME direction, we have a POSITIVE CORRELATION.
If the change is in OPPOSITE direction, we have NEGATIVE correlation
If we can’t see any form of change between the two variables… then we have a ZERO or NO CORRELATION.
A scatter plot is a graph that we can use to visualize correlation.
A correlation coefficient is a number (between -1 and +1) to quantify correlation
A scatter plot is a graph that we can use to visualize correlation.
A correlation coefficient is a number (between -1 and +1) to quantify correlation
TYPES OF CORRELATION
1)Pearson’s correlation: used for normally distributed data.
2)Spearman’s ranking: used for non normally distributed data.
3)Kendall’s tau: used when we have at least an ordinal variable
4)Cramer’s V : used for two categorical variables
1)Pearson’s correlation: used for normally distributed data.
2)Spearman’s ranking: used for non normally distributed data.
3)Kendall’s tau: used when we have at least an ordinal variable
4)Cramer’s V : used for two categorical variables
COVARIANCE
Unlike correlation that measure the strength and direction of association, covariance only measures the direction of association.
We can have positive covariance or negative covariance.
At positive covariance, both variables tend to be high and low at the same time
Unlike correlation that measure the strength and direction of association, covariance only measures the direction of association.
We can have positive covariance or negative covariance.
At positive covariance, both variables tend to be high and low at the same time
For negative covariance, one variable tends to be high while the other tends to be low.
Unlike correlation whose coefficient is usually between -1 and +1, the covariance can take values between -∞ and + ∞.
The higher the number, the higher the covariance
Unlike correlation whose coefficient is usually between -1 and +1, the covariance can take values between -∞ and + ∞.
The higher the number, the higher the covariance
REGRESSION
In the simplest form, regression is used to predict one variable from other variable(s).
In it base form, it has one dependent, one independent, one slope and one intercept…. We can this the SIMPLE LINEAR REGRESSION EQUATION
In the simplest form, regression is used to predict one variable from other variable(s).
In it base form, it has one dependent, one independent, one slope and one intercept…. We can this the SIMPLE LINEAR REGRESSION EQUATION
When you have one dependent variable and several independent variables, then we have a MULTIPLE LINEAR REGRESSION.
Other type of regression includes
Polynomial regression
Ridge regression
Logistic regression
Ordinal regression
Etc
Other type of regression includes
Polynomial regression
Ridge regression
Logistic regression
Ordinal regression
Etc
Now…. There is a relationship between the Pearson’s correlation coefficient (r) and the slope of a simple linear regression (b)
Here you go ⬇️😴
Here you go ⬇️😴
If you enjoyed this mini thread and also learnt something new, don’t forget to like this, retweet and follow me for more.
I also have a YouTube channel where I teach statistics, you can check it out ⬇️
youtube.com
I also have a YouTube channel where I teach statistics, you can check it out ⬇️
youtube.com
Loading suggestions...