This first part of this folder contains 6 files based on 73 
samples on Leukemia measured on the same set of 7129 genes. 


ALL_vs_AML_train_set_38_matrix.txt: trianing data matrix. The 
first 27 columns correspond to acute lymphocytic leukemia (ALL), 
the remaining 11 columns correspond to actute mylogous leukemia 
(AML).

ALL_vs_AML_train_set_38_sorted.res: original file containing the training data 
matrix downloaded from http://www.broad.mit.edu/cgi-bin/cancer/datasets.cgi. 
See more info there. 

Leuk_ALL_AML_test.cls: original file containing the 
test data matrix downloaded from the address above. 

Leuk_ALL_AML_test_matrix.txt: test data matrix. The first 21 columns 
correspond to ALL, the remaining 14 columns correspond to AML. 



The second part of this folder contains 3 data files corresponding to the 
above Leukemia data. One difference is that we now have 72 samples rather 
than 73. The second difference is that the data set has been pre-feature 
selected and standardized including an logarithmic (base 10) transform. The 
number of features is now only of 3571. See details in Dettling's paper. 

The 3 files are: 

leukemia.rda, downloaded from http://stat.ethz.ch/~dettling/bagboost.html.
leukemia.x.txt, data matrix (3571 by 72) ready for use in Matlab.   
leukemia.y.txt, vector of class index (72 by 1). 

Relevant papers: Golub et al. (1999). Dettling (2004), Bioinformatics.