LINCO <data>,<matrix_of_coefficients>
computes and saves linear combinations of variables in <data> into
the same data (file). <matrix_of_coefficients> is a standard MAT
file with following properties:
The names of variables to be used in linear combinations are the
row labels (8 first characters will be matched) of the matrix.
An additive constant must be indicated by row label 'constant' or
'-'.
The names of the linear combinations are the column labels of the
matrix. New variables (of 4 bytes) are created when necessary.
Lagged variables may be used by giving in the matrix an extra
column with label '#lag' and with integer elements. In this column
negative values refer to lags and positive to leads.
LINCO will be an important auxiliary operation when computing
various derived variables in linear models, multivariate analysis, etc.
Several statistical modules generate their own matrices of coeffi-
cients to be used in LINCO. For example, LINREG gives the regression
coefficients (REG.M) to be used in computing predicted values of
the model (See also OUTCNTRL?).
Example: 11 * w=1/5
12 *MATRIX MOV5
13 */// %2 #lag
14 *%1 w -2
15 *%1 w -1
16 *%1 w 0
17 *%1 w 1
18 *%1 w 2
19 *
20 *MAT SAVE MOV5
21 *LINCO DATA1,MOV5(Temp,Ave5)
LINCO on line 21 computes and saves a five term moving average 'Ave5'
of variable 'Temp' in data 'DATA1'. The matrix of coefficients 'MOV5'
has to be saved by MAT SAVE MOV5 on line 20 before activating LINCO.
The matrix of coefficients may have replaceable parameters %1,%2,...
as column and row labels, too. In this example MOV5(Temp,Ave5) means
that %1 will be replaced by 'Temp' and %2 by 'Ave5'. Hence, the same
matrix may be used for different applications.
V = More information on transformations (VAR)
O = Output matrices (giving coefficients)
S = More information on statistical operations
Help System (web edition)