A SAS analyst uses PROC TEXTMINE to analyze 50,000 product reviews. After parsing, the term-document matrix has 25,000 unique terms. The analyst wants to reduce dimensionality using SVD. Which PROC TEXTMINE option controls the number of SVD dimensions (latent topics) retained?
Select an answer to reveal the explanation.
Short Explanation
Here's the deal — you'll use NDIM= in the SVD statement because it tells PROC TEXTMINE how many singular values and vectors to compute, which is the number of latent topics you retain. For a large review corpus, you'd commonly pick something in the 50-200 range.
Full Explanation
NDIM= in the SVD statement is correct because PROC TEXTMINE uses it to specify how many singular values/vectors to compute and retain as latent semantic topics; large corpora often use 50-200. NFACTOR= in the PARSE statement controlling linguistic factors is wrong because it is not a PROC TEXTMINE PARSE option. MAXDIM= in the PROC TEXTMINE statement limiting total matrix dimensions is wrong because it is not a PROC TEXTMINE option. K= in the CLUSTER statement controlling the number of text clusters is wrong because it controls clustering, not SVD dimensions.