NOTE: 1. I changed CONTRIB and ESTIMATE to $ thousands. 2. I restricted the predictor variables to TURNOVER, ELIGIBLE, MATC, SALARY, SUSAN and VEST. ASSIGNMENT: 1. Calculate the underwriter's errors as (CONTRIB - ESTIMATE), take the absolute values of the errors and then find their mean. 2. Now using your pool of six predictors, build the best three predictor model build you can each of your three should have a p-val 3.05). 3. Calculate the residuals (Y, -)= e for your model, then find their absolute values and find their mean. 4. Compare the underwriter's average absolute error w/your regr model's mean average absolute error. 5. For observation #1, compare the underwriter's error to your model's error (for regr model, error; = residual; = (Y, - ) DATA SET GROUP ELIGIBLE FAILSAFE CONTRIB TURNOVER VEST MATCH 36675 1 14.00 691 0 25 63733 0 10.00 33 1 0 50 25560 1 8.00 21 1 0 50 177970 0 10.00 67 1 0 25 86873 1 10.00 47 1 0 50 39051 0 10.00 12 1 0 O 131449 1 12.00 85 1 0 25 30711 0 10.00 401 0 50 13691 0 10.00 30 1 1 0 49587 1 10.00 63 1 0 25 37898 0 10.00 25 1 1 100 41686 1 0.10 30 0 0 100 107657 0 14.00 21 1 0 50 39811 0 14.00 29 1 0 50 148274 0 14.00 991 0 0 SALARY SUSAN ESTIMATE 26296.3 75432 14133.7 50000 24000.0 45000 36833.3 235000 41140.1 146965 66463.8 95000 25929.3 155045 17198.7 54392 28788.6 32124.4 100000 35288.1 86000 23266.7 56000 54798.0 115650 18617.9 50000 25516.6 252316 0 on the intention na NOTE: 1. I changed CONTRIB and ESTIMATE to $ thousands. 2. I restricted the predictor variables to TURNOVER, ELIGIBLE, MATC, SALARY, SUSAN and VEST. ASSIGNMENT: 1. Calculate the underwriter's errors as (CONTRIB - ESTIMATE), take the absolute values of the errors and then find their mean. 2. Now using your pool of six predictors, build the best three predictor model build you can each of your three should have a p-val 3.05). 3. Calculate the residuals (Y, -)= e for your model, then find their absolute values and find their mean. 4. Compare the underwriter's average absolute error w/your regr model's mean average absolute error. 5. For observation #1, compare the underwriter's error to your model's error (for regr model, error; = residual; = (Y, - ) DATA SET GROUP ELIGIBLE FAILSAFE CONTRIB TURNOVER VEST MATCH 36675 1 14.00 691 0 25 63733 0 10.00 33 1 0 50 25560 1 8.00 21 1 0 50 177970 0 10.00 67 1 0 25 86873 1 10.00 47 1 0 50 39051 0 10.00 12 1 0 O 131449 1 12.00 85 1 0 25 30711 0 10.00 401 0 50 13691 0 10.00 30 1 1 0 49587 1 10.00 63 1 0 25 37898 0 10.00 25 1 1 100 41686 1 0.10 30 0 0 100 107657 0 14.00 21 1 0 50 39811 0 14.00 29 1 0 50 148274 0 14.00 991 0 0 SALARY SUSAN ESTIMATE 26296.3 75432 14133.7 50000 24000.0 45000 36833.3 235000 41140.1 146965 66463.8 95000 25929.3 155045 17198.7 54392 28788.6 32124.4 100000 35288.1 86000 23266.7 56000 54798.0 115650 18617.9 50000 25516.6 252316 0 on the intention na
