Semi-Supervised transduction

 

SSTransduction.exe utility performs a transductive inference on a joint labeled and unlabelled dataset. The input is a set of unlabelled vectors in the form of a sparse trainset (“.sts”) file. The vector parameters in the input file are assumed to be 0.0 for ‘unlabelled’ and positive integers 1..N for classess 1 to N. The result is written to the output file with the same vector ordering.

 

usage: SSTransduction.exe

-i:Input-SparseTrainSet-Data (default:’’)

-o:Output-SparseTrainSet-Data (default:’’)

 

Example:

 

SSTransduction.exe –i:input_dataset.sts –o:output_dataset.sts