| dc.description.abstract | Finding rules from data has been an active research field in Artificial Intelligence.
Active research was done in frequent itemset mining, which ultimately led to rule
generation. Many algorithms propose suggestions to overcome various problems.
Many optimization approaches focus on minimizing the creation of frequent
itemsets and scan times to trim data. The problem of finding association rules was
first introduced in an algorithm called AIS’s proposal for mining association rules.
The effectiveness of the algorithm is tested by applying a data set obtained from
large retail companies. Apriori is a Breadth First Search Algorithm (BFS) that can
generate k+1-itemsets based on frequent k-itemsets. The frequency of an item set is
calculated by counting its occurrence in each transaction. The apriori algorithm is
a data mining algorithm that is used to analyze databases based on their frequency,
based on an association rule learning system. It is designed to be applied to datasets
with transactions to increase database priority. With a fixed minimum number of
supports, Apriori scans individual databases and finds databases with high
occurrence frequency. After that process, get the minimum amount of confidence
from this set of items, the confidence value is expected to have a very good level of
accuracy to produce a good enough model. Fuzzy Mamdani is here to support a
method that can improve the accuracy of trust values and produce a model with the
hope of being able to perform itemset elimination based on the a priori algorithm.
In this study the application of Fuzzy mamdani uses 27 Rules as the formation of
Fuzzy Associations, 3 functional and 3 variables, in this case the required
processing time is 34.44 seconds for the execution of 407,700 data records. | en_US |
