Forum: open_discussion
Monitor Forum | 
Start New Thread
| MSDem 0.0.3.1 released [ Reply ] By: Marcus Wurzer on 2018-03-09 11:30 | [forum:45746] |
|
MSDem 0.0.3.1 is now available for download. Changes in MSDem 0.0.3.1: - The internal lt.axmx()-function now checks if the ax- and mx-values in the state space are compatible. If not, the ax-values are auto-adjusted and a warning message is printed ("ax has been adjusted to match mx values"). This happens if the condition ax = 1/mx is not met for the highest age group (with a tolerance of 1e-6 when comparing the values). Previously, an incorrect specification could lead to implausible nSx-values (outside the 0-1 range) and an error was thrown. - Fixed bug in internal function eapr() that caused the following error: "Error in dimnames(x) <- dn: length of 'dimnames'[2] not equal to array extent". - The state.space()-function now includes a new argument "by" that is used to specify the period length, and consequently, the definition of the age groups. Currently, it is possible to choose period lengths of 1 or 5 years (with 5 as default), and thus, one- or five-year age groups. Please note that msproj() doesn't provide the functionality to deal with period lengths of 1 year yet. - Because of the new "by" argument, the definitions of the "period" and "age" arguments have changed. Now, users only provide the lowest and highest values, respectively. The increment of the generated sequence is taken from the "by" argument. - The argument used to specify how the nSx's are computed within the msproj()-function is now called "nSx" (with uppercase "S") instead of "nsx". |
|
