Singular Value Decomposition analysis

when called separately, requires as input a file containing a matrix of occurrences of the vocabulary items (columns) by the context units identified for a particular text (columns), which is optionally generated by HAMLET Joint Frequencies, or document profiles by words created in the course of applying Latent Dirichlet Allocation and correspondence analysis, or as matrices of the same categories (column variables) for a series or separate texts (row variables) selected using the fast-track procedure for multiple file comparisons.

The document profiles are considered as ‘scores’, which are first normalised to unit length, for convenience, to enable them to be plotted as vectors through the configuration of points representing the word tokens.

When a valid file name is found, the vocabulary items available are listed as follows: