For my company AutoZone which is my left hand side I will be using 36 observations found from revenue 2006 to 2015 quarterly. The purposes for the regression analysis is to see if data has problems, test to make sure for the problem by visually and statistically, see the consequences for the problem, fixes that are available, and finally fix the problem to get better result in forecasting.
Next using the BLUE which is minimizes the sum of square errors. This is a fitting the curve or ordinary least square and get rid of value in order to have best fit. Ordinary least square is called the estimator. Best meaning the variance of variable is small as possible. Linear is a straight line of variable not squared. Unbiased the estimate of actual value of residual very close together …show more content…
For every unit increase in the consumer price index, AutoZone increased by 6928 The higher the gas price people drive less so that will result in less auto parts. For every gallon increase in the Gas Price, AutoZone increase by 347. The less unit labor cost the more unit are produced for auto parts. For every unit increase in the unit labor cost, AutoZone increased by 3140. My signs of the equation and output are matching in the correlation matrix.
I will be running time series and ACF to see if there is seasonality. Since I see on the time series and on ACF there is seasonality I will be using dummy variables along with my RHS variables to run my regression so this would reduce my error and not give me a residual. The dummy variable will determine how dependent variables related to qualitative independent variables. This graph tells that it has seasonality from quarter 4, 8, 12, 16 according to ACF. Dummy is reduce my error unexplained square is higher than it should be but including the seasonality this reduce unexplained sum of
Next using the BLUE which is minimizes the sum of square errors. This is a fitting the curve or ordinary least square and get rid of value in order to have best fit. Ordinary least square is called the estimator. Best meaning the variance of variable is small as possible. Linear is a straight line of variable not squared. Unbiased the estimate of actual value of residual very close together …show more content…
For every unit increase in the consumer price index, AutoZone increased by 6928 The higher the gas price people drive less so that will result in less auto parts. For every gallon increase in the Gas Price, AutoZone increase by 347. The less unit labor cost the more unit are produced for auto parts. For every unit increase in the unit labor cost, AutoZone increased by 3140. My signs of the equation and output are matching in the correlation matrix.
I will be running time series and ACF to see if there is seasonality. Since I see on the time series and on ACF there is seasonality I will be using dummy variables along with my RHS variables to run my regression so this would reduce my error and not give me a residual. The dummy variable will determine how dependent variables related to qualitative independent variables. This graph tells that it has seasonality from quarter 4, 8, 12, 16 according to ACF. Dummy is reduce my error unexplained square is higher than it should be but including the seasonality this reduce unexplained sum of