Category: Functions
| Q | |
| Q | Can I compute the logarithmic return from a time series without any intermediate calculations? |
| Q | |
| Q | How can I use NumXL to compute the correlation matrix for 2 time series? |
| Q | |
| Q | How do I use differencing to obtain a stationary time series? |
| Q | |
| Q | The calibration with Excel 2010 Solver does not give optimal coefficients. |
| Q | |
| Q | When I try to use the calibration icon, an error message pops stating that solver is not found. |
| Q | |
| Q | |
| Q |
Why does the calibration function in NumXL shows the constraint's rhs of 0.9999? |
| A | When you click on the "Calibration" icon in the toolbar (or menu), NumXL prepare all inputs of the solver using the selected model, and add one constraint (As shown below):
The constraint is added to make sure the solver search only for a stable model (e.g. invertible, characteristics root outside unit-circle, etc.). NumXL has functions for the different models that checks the stability of the model and returns either 0/1. Ideally, the constraing above should state something like ($J$43=1), but due to an issue in the Excel Solver using VBA, we introduced a work-around to overcome situations where the first constraint is missing or dropped. |
