# Correlation Coefficients

## Correlation Coefficients • What are correlation coefficients? In simple words, Correlation coefficients are used to show a linear relationship between two
different variables. In our dataset whenever we want to check how much one column(variable x)
is dependent on another column (variable y) then we use " Correlation Coefficients ". there
are two types of correlation.

1 - positive correlation
2 - negative correlation. correlation coefficients will give you a value between -1 to 1 (negative one to positive one)

1 - Positive Correlation
when the correlation coefficient is greater than 0 it considers a positive correlation.
in simple words, positive correlation means when one feature (variable x) starts increasing
then another feature (variable y) also starts increasing and when one feature (variable x)
starts decreasing then another feature (variable y) also starts decreasing.

Example:-
suppose we have two columns in our dataset selling and profit.

``````| variable x | variable y |
| --- | --- |
| selling | profit |
``````

if we find a correlation coefficient between this two-column it will give Positive Correlation
because we know that if selling will go high then the profit will also go high and if
selling will go low then the profit will also go low. Positive correlation will give you a value between 0 to 1

2 - Negative Correlation
when the correlation coefficient is lower than 0 it considers a negative correlation.
in simple words, negative correlation means when one feature (variable x) starts increasing
then another feature (variable y) starts decreasing and when one feature (variable x)
starts decreasing then another feature (variable y) starts increasing.

Example:-
suppose we have two columns in our dataset price and selling rate.

``````| variable x | variable y |
| --- | --- |
| price | selling rate |
``````

if we find a correlation coefficient between this two-column it will give Negative Correlation
because we know that if the price will go high then the selling rate will go low and if
price will go low then the selling rate will go high. Negative correlation will give you a value between 0 to -1

3 Likes

Does a correlation coefficient have any way of indicating if the relationship is linear or not?