Skip to content

Latest commit

 

History

History
24 lines (20 loc) · 1.08 KB

File metadata and controls

24 lines (20 loc) · 1.08 KB

#flashcards/stats

Correlation

Correlation is a measure of strength in a linear relationship $r$ is the Correlation coefficient, which describes the amount of Correlation in a data set. Formula for correlation coefficient::$$r=\frac{\sum_{i=1}^n(x_i-\bar{x})(y_i-\bar{y})}{(n-1)s_xs_y}$$

Where $\bar{x}$ is the [[Measures of central tendency|mean]] and $s_x$ is the [[Frequency distributions#Standard deviation|standard deviation]]

for [[Jointly distributed random variables]],$$\rho_{x,y}=\frac{\sigma_{xy}}{\sigma_x\ \sigma_y}$$Where $\sigma_x$ is the variance of [[random variables|random variable]] x and $\sigma_{xy}$ is the [[Jointly distributed random variables|covariance]] of X and Y

r or $\rho$ > 0: +ve correlation r or $\rho$ < 0: -ve correlation r or $\rho=0$ : No correlation

Properties of $r$

?

  1. $-1\le r\le1$
  2. $y_i=a+bx_i$ then r=1 (b>0)
  3. $y_i=a+bx_i$ then r=-1 (b<0)
  4. Correlation does not change if constants change(remain +ve) in a [[Linear transformation]] $x_i$,$y_i$ -> $a+bx_i$,$c+dy_i$ r remains constant