momentd#

Purpose#

Computes a moment (x’x) matrix from a GAUSS dataset.

Format#

m = momentd(dataset, vars)#
Parameters:

  • dataset (string) – name of dataset.

  • vars (Kx1 string array or Kx1 numeric vector or string) –

    Type

    Contents

    Kx1 string array

    names of variables

    Kx1 numeric vector

    indices of columns

    string

    Formula string e.g. "PAY + WT" or ". - 1" (include all variables besides intercept).

    These can be any size subset of the variables in the dataset, and can be in any order. If a scalar 0 is passed, all columns of the dataset will be used.

Returns:

m (MxM matrix) –

where \(M = K + \_\_con\), the moment matrix constructed by calculating \(X'X\) where X is the data, with or without a constant vector of ones.

Error handling is controlled by the low order bit of the trap flag.

trap 0:

terminate with error message

trap 1:

return scalar error code in m

33

too many missings

34

file not found

Global Input#

__con:

(scalar), default 1.

1

a constant term will be added.

0

no constant term will be added.
__miss:

(scalar), default 0.

0

there are no missing values (fastest).

1

do listwise deletion; drop an observation if any missings occur in it.

2

do pairwise deletion; this is equivalent to setting missings to 0 when calculating m.
__row:

(scalar), the number of rows to read per iteration of the read loop, default 0.

If 0, the number of rows will be calculated internally.

If you get an Insufficient memory error, or you want the rounding to be exactly the same between runs, you can set the number of rows to read before calling momentd().

Examples#

Using indices of columns#

fname = getGAUSShome() $+ "examples/freqdata.dat";

// Calculate statistics on variables in dataset: PAY and WT
// Specify the index of PAY and WT
vars = 2|4;

// Find moment
m = momentd(fname, vars);

print  m;

After the above code,

400.00000        787.00000        587.98000
787.00000        1805.0000        1161.1400
587.98000        1161.1400        900.38540

Using names of variables#

// Get filename
fname = getGAUSShome() $+ "examples/freqdata.dat";

// Calculate statistics on variables in dataset: PAY and WT
// Define the names string array of PAY and WT
string vars = {"PAY", "WT"};

// Find moment
m = momentd(fname, vars );
print  m;

After the above code,

400.00000        787.00000        587.98000
787.00000        1805.0000        1161.1400
587.98000        1161.1400        900.38540

Using formula string#

// Get filename
fname = getGAUSShome() $+ "examples/freqdata.dat";

// Define the formula for PAY and WT
formula_str = "PAY + WT";

// Calculate statistics on variables in dataset: PAY and WT
m = momentd(fname, formula_str);
print  m;

After the above code,

400.00000        787.00000        587.98000
787.00000        1805.0000        1161.1400
587.98000        1161.1400        900.38540

Remarks#

  • The supported dataset types are CSV, Excel, HDF5, GAUSS Matrix (FMT), GAUSS Dataset (DAT), Stata (DTA) and SAS (SAS7BDAT, SAS7BCAT).

  • Character vectors are supported for backward compatibility, but it has been deprecated.

Source#

momentd.src

See also#

See also

Formula String

On this page