canoalberto / ur-caim Goto Github PK

great code thanks
only
1
can you share some data file to test how it works , to compare my calculation results with your calculation results to make sure I use it in correct way
2
may be know somebody else implementation not using weka , but only plain java code?

may you pls help to find what data sets where used since link from paper is broken
since
link from paper page 7
https://helvia.uco.es/xmlui/bitstream/handle/10396/13003/article.pdf?sequence=1&isAllowed=y
is broken
The data sets description along with their partitions, the ur-CAIM
source code and WEKA plugin, the experimental settings and results
for all data sets and algorithms are fully described and publicly
available to facilitate the replicability of the experiments and future
comparisons at the website:
http://www.uco.es/grupos/kdis/wiki/ur-CAIM

equations 8,9,13 in paper you use log2
but in code
my guess you tried to implement logarithm base change like
https://www.khanacademy.org/math/algebra2/x2ec2f6f830c9fb89:logs/x2ec2f6f830c9fb89:change-of-base/a/logarithm-change-of-base-rule-intro

maybe you meant using only log10
Math.log10(Pr / Pir) / Math.log10(2)

changing to Math.log10(Pr / Pir) / Math.log10(2)
leads to calculation results change ?

java code reference
urCAIM.java
equations 8,9,13
if(Pir != 0)
{
H += Pir * Math.log10(1.0 / Pir) / Math.log(2);
I += (1-Pi) Pir Math.log10(Pir / (Pi * Pr)) / Math.log(2);
INFO += Pir * Math.log10(Pr / Pir) / Math.log(2);
}

227-232 lines

Thanks for help

GREAT ALGORITHM
BUT DO HAVE IT IN PYTHON ?