How to compute histograms using weka -

given dataset 23 points spread out on 6 dimensions, in first part of exercise should following, , stuck on second half of this:

• compute first step of clique algorithm (detection of dense cells). use 3 equal intervals per dimension in domain 0..100,and consider cell dense if contains @ least 5 objects.

now trivial , matter of counting. next part asks following though:

• identify way compute above clique result using functions of weka provided in tabs of preprocess, classify , cluster , or associate . hint : 2 tabs needed.

i've been trying on hour now, can't seem anywhere near solution here. if has hint, or maybe useful tutorial gives me little more insight weka appreciated!

i assuming have 23 instances (rows) , 6 attributes (dimensions)

use 3 equal intervals per dimension

use pre-process tab discretize data 3 equal bins. see image or command line. use 3 bins intervals. may choose change useequalfrequency false , true , try again. think true may give better results.

weka.filters.unsupervised.attribute.discretize -b 3 -m -1.0 -r first-last

after cluster data. give show near instances. since find dense cells. think som may appropriate.

a cell dense if contains @ least 5 objects.

you have 23 instances. therefore try 2x2=4 cluster centers, go 2x3=6,2x4=8 , 3x3=9. if data points near. of cluster centers should hold 5 instances no matter how many cluster centers choose.