THIS INFORMATION IS OBSOLETE AND IS PROVIDED ONLY FOR ITS HISTORICAL VALUE

by E. R. Malinowski, P. H. Weiner, P. T. Funke and A. Levinstone, Department of Chemistry and Chemical Engineering, Stevens Institute of Technology, Hoboken, New Jersey 07030 and D. G. Howery, J. M. Soroka and R. B. Selzer, Department of Chemistry, Brooklyn College of CUNY, Brooklyn, New York 11210

FACTANAL allows one to carry out principal component target-transformation factor analyses of complete matrices of chemical data. Using a Jacobi diagonalization method, the program computes the eigenvalues(abstract factors)forthe correlation/covariance matrix formed from the data matrix.The data matrix is reproduced using combinationsinvolvingsuccessivelysmaller eigenvalues. The program performs least-squares target transformations of test vectors related to the physical and structural parameters of the row designees of the matrix (and of the column designees using the transposed matrix). The target-transformation capability is not available in standard factor analysis programs. The matrix can be calculated for a given number of factors employing combinations of columns and of rows from the data matrix and combinations of specified complete physical and structural test vectors. A large number of useful printing and tabulational options are included. Sample output for a test problem is explained in the documentation.

FORTRAN IV (IBM 370/165) Lines of Code: 2610 Recommended Citation: E. R. Malinowski et al., QCPE 11, 320 (1976).