Fig.11A: Triple Matrix Data Patterns before Optimization (n=44)
Fig.11B Data Patterns after Removal of non informative Columns (n=23)
- Displayed are the disease classification masks
for reference and risk patients ("disease signatures") as well as the
patient classification masks ("patient signatures") for the
first 10 control and myocardial infarction risk patients
containing 44 data columns each (upper panel).
- Data columns containing (0)=unchanged as most frequent triple
matrix character in both classification categories do not contain
discriminatory information and are therefore removed (lower panel),
thus leaving 23 data columns for the subsequent optimization process.
- Blue colored columns represent
the finally selected most discriminatory data columns