In this paper, a novel method named as
Frequency Pattern-Knowledge Constructions (FP-KC) is
developed. This method attempts to develop Frequency Pattern
(FP) Growth data mining algorithm using several knowledge
constructions to find the association rules and minimize the
shared information ( i.e. fined frequent item set), FP-KC
combines the criteria of Principal Component Analysis (PCA)
with FP-Growth techniques. These criteria include eigenvalues,
cumulative variability and scree plot. There are several reasons
for developing the FP-Growth data mining algorithm to build up
a novel FP-KC technique that can find the association rules,
including: (a) the size of an FP-tree is typically smaller than the
size of the uncompressed data because many records in a dataset
often have a few items (b) to give the best result in the case that
all the records have the same set of items; (c) FP-Growth is an
efficient algorithm because it illustrates how a compact
representation of the transaction dataset helps to efficiently
generate frequent item sets; and (d) The run-time performance of
FP-Growth depends on the compaction factor of the dataset,
while the enhanced algorithm in Subliminal Channel (SC)
depends on both the position of a character in the alphabet and
its position in the plain rule word (i.e. rules resulting from
association rules FP-KC), with a specific function to determine
the cipher rule character. To evaluate the efficiency of the
proposed method, four case studies were used. Based on the
results, the proposed method can be considered as an efficient
technique for secure mining of association rules of partitioned
data compared with the traditional method.