Departmental Processing
Turnaround Estimation Project
| Home |
| Problem definition |
| Data collected |
| Regression model |
| Analysis |
| Stepwise |
| Final model |
| Forecast Model for Duration |
| Load Chart |
Basic Estimated Prediction/ Forecast Model for Duration:
| #ENG | RFQFORM | MODCAT | CMPLXCAT | Predicted Lower Limit @ | Predicted Mean Duration (DUR) | Predicted Upper Limit @ |
| 2 | 0 | 0 | 0 | 8.26 | 9.29 | 10.33 |
| 2 | 0 | 1 | 0 | 27.50 | 29.14 | 30.78 |
| 2 | 0 | 0 | 1 | 38.11 | 42.03 | 45.95 |
| 2 | 1 | 0 | 0 | 4.54 | 6.12 | 7.70 |
| 2 | 1 | 1 | 0 | 24.05 | 25.97 | 27.88 |
| 2 | 1 | 0 | 1 | 34.77 | 38.86 | 42.95 |
| 3 | 0 | 0 | 0 | 5.31 | 6.54 | 7.77 |
| 3 | 0 | 1 | 0 | 24.61 | 26.39 | 28.17 |
| 3 | 0 | 0 | 1 | 35.22 | 39.28 | 43.33 |
| 3 | 1 | 0 | 0 | 2.08 | 3.37 | 4.66 |
| 3 | 1 | 1 | 0 | 21.52 | 23.22 | 24.91 |
| 3 | 1 | 0 | 1 | 32.04 | 36.11 | 40.17 |
Results / Conclusion:
Because of the level of overall significance and despite the relatively benign “fit” that we achieved with our analysis, we feel that this model is a good first step in providing the Order Engineering department with a prediction or forecast model for the number of days that any particular special order should take. Based on the significant variables that are present in our final regression model equation, it became very evident that there were a finite number of possible outcomes from this equation. Since the number of engineers in the department will most likely not drop below 2 nor increase over 3 (i.e. x2 = 2 or 3) and the remaining variables were dummies standing in for a qualitative “recorded” variable, there could actually be only 12 possible solutions to this equation, as detailed in Table 2 of the Figures and Tables section of this report. This table is basically laid out in a matrix with the user finding the row whose corresponding columns match their data and then follow the row over to the mean and upper and lower “estimated” project duration from the appropriate columns on the right.