# 8 year redictor onstant ear coef 28438 14822 176933 se

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: the residuals chart MUST NOT have any patterns such as curve or wedge Scatterplot of Tornadoes vs Year 2000 1750 Tornadoes 1500 1250 1000 750 500 1950 1960 1970 1980 Year 1990 2000 2010 WE SEE THAT X AND Y HAVE A LINEAR RELATIONSHIP, WE CAN DO LINEAR REGRESSION x=year y=qty of tornadoes x predicts y he regression equation is ornadoes = - 28438 + 14.8 Year redictor onstant ear Coef -28438 14.822 = 176.933 SE Coef 2897 1.463 T -9.82 10.13 R-Sq = 65.5% P 0.000 0.000 R-Sq(adj) = 64.9% nalysis of Variance ource egression esidual Error otal DF 1 54 55 SS 3214212 1690492 4904704 MS 3214212 31305 F 102.67 P 0.000 nusual Observations bs 35 52 Year 1987 2004 Tornadoes 656.0 1819.0 Fit 1013.4 1265.4 SE Fit 25.5 41.7 Residual -357.4 553.6 St Resid -2.04R 3.22R Versus Fits (response is Tornadoes) 600 Residual 400 200 0 -200 -400 500 600 700 800 900 1000 Fitted Value 1100 1200 1300 1400 TWO BASIC TESTS PASSED, now: a. DISCUSS r‐squared r‐SQUARED IS how well the Predictor EXPLAINS the response 65% OF THE VARIATION IN THE QUANTITY OF TORNADOES IS EXPLAINED BY YEAR. b. DISCUSS THE MAXIMUM ERROR The model’s STANDARD DEVIATION is the model’s measure of error, which also called STANDARD ERROR Se comes in one unit. IT FOLLOWS THE NORMAL DISTRIBUTION SO WE want to double that (times two) so that THE ERROR APPLIES to 95% of the times. THUS, 353 is the model’s maximum error in prediction SAY Y‐HAT IS 800, then THE TRUE VALUE CAN BE ANYWHERE FROM 800‐353 to 800+353. c. SHOW MODEL AND PREDICT SOMETHING ANOTHER MODEL TUITION IN 2000 PREDICTS TUITION IN 2008 RA...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online