Stata 15 help for matrix score

[P] matrix score -- Score data from coefficient vectors

Syntax

matrix score [type] newvar = b [if] [in] [, equation(##|eqname) missval(#) replace forcezero]

where b is a 1 x p matrix.

Description

matrix score creates newvar = x_jb' (b being a row vector), where x_j is the row vector of values of the variables specified by the column names of b. The name _cons is treated as a variable equal to 1.

Options

equation(##|eqname) specifies the equation -- by either number or name -- for selecting coefficients from b to use in scoring. See [P] matrix rownames for more on equation labels with matrices.

missval(#) specifies the value to be assumed if any values are missing from the variables referred to by the coefficient vector. By default, this value is taken to be missing (.), and any missing value among the variables produces a missing score.

replace specifies that newvar already exists. Here observations not included by if exp and in range are left unchanged; that is, they are not changed to missing. Be warned that replace does not promote the storage type of the existing variable; if the variable was stored as an int, the calculated scores would be truncated to integers when stored.

forcezero specifies that, should a variable described by the column names of b not exist, the calculation treat the missing variable as if it did exist and was equal to zero for all observations. It contributes nothing to the summation. By default, a missing variable would produce an error message.

Example

Setup . sysuse auto . regress price weight mpg

Define matrix coefs equal to e(b), the coefficient vector . matrix coefs = e(b)

List the contents of coefs . mat list coefs

Create variable lc containing the linear predictions . matrix score lc = coefs

Summarize lc . summarize lc

Setup . sureg (price weight mpg) (displacement weight)

Define matrix coefs equal to e(b), the coefficient vector . matrix coefs = e(b)

List the contents of coefs . mat list coefs

Create variable lca containing the linear predictions for equation price . matrix score lca = coefs, eq(price)

Same as above command . matrix score lc1 = coefs, eq(#1)

Same as above command . matrix score lcnoeq = coefs

Create variable lcb containing the linear predictions for equation displacement . matrix score lcb = coefs, eq(displacement)

Same as above command . matrix score lc2 = coefs, eq(#2)

Summarize newly created variables . summarize lc*


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index