Skip to content

tmpinv 1.1.2 for SSC

Latest
Compare
Choose a tag to compare
@econcz econcz released this 22 Feb 10:40
a1cb0ce

The program implements a non-iterated Transaction Matrix (TM)-specific LPLS estimator for linear programming with the help of the Moore-Penrose inverse (pseudoinverse), calculated using singular value decomposition (SVD). Estimation using 2x2 to 50x50 contiguous submatrices, repeated with compensatory slack variables until NRMSE is minimized in a given number of iterations, is followed by an F-test from linear regression/t-test of mean NRMSE from a pre-simulated distribution (Monte-Carlo, 50,000 iterations with matrices consisting of normal random variates). The result is adjusted for extreme values to match the RHS via shares of estimated row/column sums if the corresponding option is specified.