UDC 004.62
THE METHOD TO COMPRESS IN INTERMEDIATE DATA
IN DISTRIBUTED MINING OF ASSOCIATION RULES
E. O. Khramshina, postgraduate student, lecturer assistant of the Department of Computational and Applied
Mathematics, RSREU, Ryazan, Russia;
orcid 0000-0002-4490-8403, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The aim of this work is to decrease memory volume for storage and transfer taken by three-dimensional
array in distributed mining of association rules. Volume decrease is achieved due to sparse array and variable
size of array element values. Array elements with non-zero values are written as a number pair: the offset
from the previous element with non-zero value and the value itself. To separate numbers, one of the bits in a
byte is used as a service one, to point the end of the value. The experiments have shown that this method allows
file size 74% less on average in comparison with the original array. Software in Java programming
language has been developed for these experiments. The compression method together with 3D2ARM association
rules algorithm can be used to develop distributed mining of association rules.
Key words: data mining, association rules mining, 3D2ARM algorithm, data structures, data transfer,
data compression, zero-length suppression method, variable size of array values.