CLASSIF1 Data Pattern Classification
fig.14: Classification of Random Number Data Set
- Fourty data records, each containing 144 data columns of computer generated
random numbers were available.
- The numbers in the various data columns had coefficients of variation
(CV=100*standard deviation/mean (%)) between 0.97% and 25.7% to mimick
flow cytometric data columns like G0/G1 peaks of DNA histograms
or antigen distribution curves in immunophenotyping.
- An alternating sequence of category 1 ("cat#1") or category 2 ("cat#2")
patient labels were assigned to the random number data records
prior to the learning phase.
- Data records 1, 5, 10, 15, 20 of "cat.1" and "cat.2" served as
unknown embedded test set, remaining hidden to the learning
Last Update: Jan 11,2017
First display: Feb 16, 2005