A new version of -moremata- is available from SSC. Type
to install the new version.
Several existing functions have been revised and there are various new functions. Some highlights:
Code:
. ssc install moremata, replace
Several existing functions have been revised and there are various new functions. Some highlights:
- new mm_density() provides a class-based system for kernel density estimation
- new mm_ddens() implements density estimation by diffusion by Botev et al. (2010)
- mm_quantile() has been rewritten; it now supports all 9 quantile definitions from Hyndman and Fan (1996); weights are supported for all definitiions; there is a new _mm_quantile() function for fast computation of quantiles if the data is already sorted. (please note: argument -altdef- in mm_quantile() and mm_iqrange() has been replaced by argument -def- that can take on values 1 to 9; altdef!=0 in the previous version is equivalent to def=6 in the new version)
- various functions now have an "underscore" variant that assumes sorted data; use these functions to save computer time if the data has already been sorted
- function mm_relrank() has been rewritten; it now has additional arguments support breaking ties and to compute non-normalized ranks
- new mm_unique() and related functions to obtain or tag unique values in a vector or unique rows in a matrix, with options to determine the order in which the result is returned
- new mm_sort() and mm_order() functions for stable sorting
- new mm_minim() function for univariate minimization without derivatives
- new mm_prod() functions to compute products of elements in a matrix
- new mm_seq() function to generate regular sequences
- new mm_diff() function for lagged differences
- new mm_group() function for creating a group index
- new mm_clip() function to clip/limit values in a matrix
- ... a history of changes can be found at https://github.com/benjann/moremata; the update covers changes since 14jul2020.
