- The category classification mask for the category#1
reference group of data records contains typically a sequence of (0)
characters because the majority (70%) of the parameter values are
located between the two percentiles thresholds
15% and 85%, that is 15% below (-) the lower percentile and
15% above (+) the upper percentile.
- Data records are classified according to the highest positional
coincidence of the data record classification masks
with the two category classification masks.
- The degree of coincidence between the data record classification mask
and the best fitting category mask is expressed by
the mask coincidence factor. The coincidence factor
is 1.00 for data records #003,013,025,006,008,024,026
despite the fact that not all triple matrix characters are (0).
Category#2 data records have all discriminatory parameters
of the data record classification masks
increased (+). Data records with normal (0) or diminished (-)
discriminatory parameter values belong therefore to category#1
data records. For this reason, parameter values (-) are counted as
hits for category#1 data records.
- Displayed are the first 10 triple matrix patterns of
category#1 and category#2 data records. Data records #001,002,
009,010,019,020,029,030,039,040 had been removed prior to the
learning phase to serve as
unknown test set.
This permits to independently assess the discriminatory potential of learned
classifiers for unknown data records.
