# MAST10011 Chapter Notes - Chapter 8: Confidence Interval, Test Statistic

## Document Summary

Regression: used for prediction, when you want to use x to predict y. Correlation: used to measure association between x and y. The average increase in y when x is increased by one. A(cid:374)d are esti(cid:373)ated, (cid:449)e ha(cid:448)e a fitted (cid:373)odel. The fitted model is used to estimate the average value of y for a given value of x. It is also used to predict a future observation of y for a given value of x. Estimating and : least squares (cid:2010) =(cid:1870)(cid:1871)(cid:3052)(cid:1871)(cid:3051), (cid:2009) =(cid:1877) (cid:2010) (cid:1876) Variances should be equal - look for trends in a residuals vs fitted plot should be randomly scattered. Should be normally distributed check with qq plot residual sum of squares: (cid:1856)(cid:2870)= ((cid:1877) (cid:2009) (cid:2010) (cid:1876))(cid:2870) error mean squared (to estimate (cid:2870)): (cid:1871)(cid:2870)= 2 (cid:2870)= (cid:2869) (cid:2870)(cid:1871)(cid:3052)(cid:2870)(cid:4666)(cid:883) (cid:1870)(cid:2870)(cid:4667) H0: = 0 (cid:894)testi(cid:374)g the utilit(cid:455) of the (cid:373)odel: is (cid:454) of a(cid:374)(cid:455) use i(cid:374) predi(cid:272)ti(cid:374)g (cid:455)? (cid:895).