Method and program for estimating uncertainty

A program for estimating uncertainty of measuring machines, comprising computer-executable steps of:

• estimating information of errors or precisions on a measuring machine to be estimated;
• deriving a covariance matrix or correlation matrix of measured values provided from the measuring machine based on the information estimated in the previous step;
• deriving eigenvalues and eigenvectors from eigenvalue decomposition of the covariance matrix or correlation matrix derived in the previous step;
• generating pseudo-measured values of the measuring machine by generating a normal random number as a coefficient of coupling for each eigenvector derived in the previous step and linearly coupling all eigenvectors, the normal random number having an expected value of 0 and a variance equal to an eigenvalue corresponding to the eigenvector;
• and
• estimating uncertainty of the measuring machine through statistical processing of the pseudo-measured values obtained in the previous step.