230
Tạp chí
Kinh tế - Kỹ thuật
Algorithm: Analysis of Data in Distributed Environment
Step1: Take the Spatial Database
Step2: Convert into the horizontally partitioned distributed database (N Number of datasets)
Step3: Calculate the support count of each database.
Step4: Calculate the negative support and confidence.
Step5: Calculate negative partial support and partial confidence.
Partial Support (PS) = 1-(X. Support - DB
minimum Support)
Partial Confidence (PC) = 1-(X. Confidence - DB
Minimum Confidence)
Step6: Add their own private key in all negative partial support and partial confidence.
Partial Support (PS) = 1-(X. support - DB minimum support Key)
Partial Confidence (PC) = 1-(X. Confidence - DB Minimum Confidence
)
Step7: Divided the negative partial support and confidence into the three different values.
Step8: Converted negative partial support, Confidence and lift values into the ASCII value and
compute the matrix Y.
Step9: Take the transpose of the matrix (Y
T
).
Step10: Exchange Y
T
into the Binary format.
Step11: Let own key matrix X
Step 12: Exchange X into binary
Step 13: Execute Ex-or among X and Y.
Step14: The matrix (Step 14) stored in associate memory.
Setp15: The resultant matrix is sanded to the protocol initiator Server.
After taking the transpose converted into the ASCII value and then, the value of the resulting
matrix M is same as the global support value, If global support value > 0 then it implies that the,
attribute value, that has been taken is globally frequent attribute, it might be locally infrequent value.
So here, this section the computed estimation of worldwide support is > 0 and it is acknowledged
as globally frequent item.
4. RESULT AND DISCUSSION
Numerous tests were directed on a 3.3 GHz Intel Processor with 8 GB main memory. Also,
tests were keep running on Windows 7 OS. All calculations utilized as a part of analyses were
composed in the MATLAB 2014.We performed the simulation for the generation of rules on three
different support and confidence value which is useful to show the performance of the proposed
work. We ingress our proposed technique in three spatial datasets: Delhi, Bhubaneswar and
Bangalore. Qualities of the datasets are appeared in Table 1 in which, second segment demonstrates
the quantity of records in each dataset and third segment demonstrates the quantity of highlights
for each dataset.
Table 1. Datasets Characteristics
Dataset
Number of records
Number of features
Delhi
96
5
Bhubaneswar
225
18
Bangalore
2178
4
