MGCR 271 Study Guide - Final Guide: John Tukey, Standard Deviation, Box Plot

• School: McGill University
• Department: Management Core
• Course: MGCR 271
• Professor: Kenneth Mac Kenzie

Document Summary

Slope of lrsl: b=r(sy/sx) r2 compares prediction error of lrsl; gives % of correlation explained by. Lsrl (how successfully the regression explains the response) Entering a pair of numbers (x,y): x std y m+ X 2nd f sto frequency m+ Outliers w/ both methods: hinges= tukey method, the median of upper and lower fence, q1/q3= interpolation method, 25% and 75% 1 = perfect positive correlation, -1 = perfect negative correlation. Outliers: z-score criterion: more or less than 3 std dev. from the mean. Calculate q1, q3 as the median of the lower and upper fence. Outlier if x is greater or less than: P% of the data falls below n: p% x (n+1) = ni, p%=ni+ decimal(ni+1-ni) Random variables: continuous use density curve, discrete use pdf (probability distribution function) table. Minimizes the distance of the squares of the residuals. If n is odd, m= middle number. If n is even, m = average of.