AutomobiledesignBasicEconometricsCaseAnshatSinghalB09070
BackgroundAutomobile designers worried about the gasoline mileage a car would giveWould it result in violation of Corporate Average fuel Efficiency regulationsWeight of the car was the main concernThe kind of factors that would affect were like a BLACK BOX to themThe Goal is to get an equation and predict the mileage.
OverviewText area
Lets begin…GPM City= .00943234+ .00001 Weight (lb)
Change of scale of parametersText areaGP1000M City= 9.43234+ .01362 Weight (lb)
Residuals
AnalysisGP1000MCity= 9.43234+ .01362 Weight (lb)For 4000 lb => 63.8 GP1000MBy 95% confidence interval =>55.3-72.5 GP1000MCost at $1.2 per gallon => 66.36 – 87.00 $/1000M
Correlations
Scatter plot
Do not forget the powerGP1000M City= 11.7 + .0089 Weight (lb) +.088 HorsepowerR2 from 77% to 84 %Horse power capture one third of residual variation with t statistic = 7.29
Scatter plot/ Residual plotDiscreetness of responseskewnessin the residual has reduced
Weight v. HorsepowerHigh Correlation
Weight is the real problem
HP/Pound by weightT statistic of weight is higherR2 is good at 84%
Variation Explained
Final..
ConclusionWeightand Horsepower are the important factorsPowerto weight gives the better equationPrediction Interval[57.3-71.3]GP1000M=> [14.0-17.5] MPG
Thank you

Automobile design