Linear Regression Flashcards
Linear regression equation (true)
yi = B0+B1xi +Ei
Linear regression equation (Fitted)
yHat = b0+b1x
b0
the intercept
b1
the slope (rise / run)
x
given value of x
Correlation
*denoted as r, measures the strength and direction of the line
*between -1 and 1. positive value = rising line, negative = falling line
*r = 0 = no linear relation between X and Y
*closer R is from 1 or -1 the stronger the relation
r^2
the variation! if you are given a % in a linear regression problem it is r^2
slope equation
b1 = r * (Sy/Sx)
Intercept equation
b0 = yBar - b1(xBar)
residual
the about of variation we could not predict. Sum of the residuals will equal 0
residual = y - yHat = observed y @ given x - predicted y @ given x
Se
different from SE, this is the standard deviation of residuals!
confidence interval for slope
b1+- t* (Seb1)
confidence interval form mean value of Y
yHat +- t* (Seu)