SDNOMINAL Test and define nominal data representation Display nominal feature information SDNOMINAL(A) SDNOMINAL(P) Test if numerical representation of nominal data is identical I=SDNOMINAL(A,B) or I=SDNOMINAL(A,P) Update data matrix of A with nominal representation in B A2=SDNOMINAL(A,'from',B) % A2 and B are comparable Join nominal representations in A and B and update both [A3,B3]=SDNOMINAL(A,'join',B) % A3 and B3 are comparable Remove nominal information from data set or pipeline A4=SDNOMINAL(A,'remove') INPUTS A,B sddata sets P trained pipeline OUTPUTS I indicator 1 if both sets are comparable, 0 if not A2 sddata set A with nominal rep. compatible with B A3,B3 sddata sets A and B with joined nom.rep. (classifier trained on A3 may be used on B3) A4 data set without nominal information OPTIONS 'nodisplay' Do not show comments 'from',B pass nominal rep. from B to A (update data matrix) 'join',B join nominal rep. in A and B (update both data matrices). 'remove',A remove nominal rep.meta-data from A (unchanged data matrix) DESCRIPTION When nominal features are present in the data set, their numerical representation is defined by featlist sdlist of nominal categories. To use the same classifier, data sets must share the same nominal representation. SDNOMINAL command allows one to check if two data sets share the same numerical representation of nominal data, update representation based on other set or join two nominal representations and update both sets. EXAMPLE I=sdnominal(A,B); % if I==0, classifier trained on A cannot be used on B SEE ALSO ISNOMINAL
sdnominal
is referenced in examples: