<![CDATA[The popular association rule algorithms are a priori and fp-growth, these two algorithms are very familiar among data mining researchers, however there are several weaknesses found in the association rule algorithm, including scanning the dataset for a long time in the process of searching for itemset frequencies, the use of large memory and the resulting base rules are sometimes less than optimal. In this research, the author compared the fp-growth and TPQ-apriori algorithms to analyze the base rule results of the two algorithms. TPQ-Apriori is an algorithm resulting from the development of the apriori algorithm, where the performance of the TPQ-Apriori algorithm is better than the traditional apriori algorithm in terms of the dataset scanning process in searching for itemset frequencies. For experiments, the fp-growth algorithm used the rapidminer tool while the TPQ-apriori algorithm used an application program that was built by ourselves. Meanwhile, the dataset used is sales data on CV. Charandita Kusuma NTB which has been uploaded to the Kaggle site. The base rules testing results are from the overall rule testing results with the CV sales dataset. Charandita Kusuma NTB can draw a conclusion that the larger the dataset to be processed, the more optimal the results will be if using the fp-growth rapidminer algorithm, but it is not optimal if the dataset to be processed is a small dataset. Some rules do not appear in the fp-growth algorithm with the rapidminer tool. Meanwhile, the TPQ-Apriori algorithm that has been developed is able to produce optimal rules for both large datasets and small datasets.]]>
Copyrights © 2024
