MARK Version 10.1 posted
Posted: Sat Apr 08, 2023 10:27 amThe following updates have been added to the current version of MARK.
March, 2023. Version 10.1.
297. I have made more changes to the Fortran code to incorporate useful features that were unavailable when I started writing MARK. The main change that users will notice is that you can now have induvial covariate names > 10 characters. There is no limit, although I suggest less than 30. The design matrix parameters under the Properties window allow 60 characters in a cell, but processing is faster if you reduce this default.
April, 2023.
298. A function to sweep the design matrix of repetitive rows has been added under the File menu choice when the design matrix is shown on the screen. This function is useful when there are many rows to produce estimates of the same real parameter. By sweeping the design matrix of these repetitive rows, the number of real parameters in the output is reduce, which also reduces the size of the model output. A wise user of this function should always check that the swept model produces the same -2log likelihood as the original model.
299. A dash of color has been added to the output screen where model optimization is taking place — just an eye opener unless a RED error message appears.
More details are provided in the MARK help file.
Gary
March, 2023. Version 10.1.
297. I have made more changes to the Fortran code to incorporate useful features that were unavailable when I started writing MARK. The main change that users will notice is that you can now have induvial covariate names > 10 characters. There is no limit, although I suggest less than 30. The design matrix parameters under the Properties window allow 60 characters in a cell, but processing is faster if you reduce this default.
April, 2023.
298. A function to sweep the design matrix of repetitive rows has been added under the File menu choice when the design matrix is shown on the screen. This function is useful when there are many rows to produce estimates of the same real parameter. By sweeping the design matrix of these repetitive rows, the number of real parameters in the output is reduce, which also reduces the size of the model output. A wise user of this function should always check that the swept model produces the same -2log likelihood as the original model.
299. A dash of color has been added to the output screen where model optimization is taking place — just an eye opener unless a RED error message appears.
More details are provided in the MARK help file.
Gary