DATA MINING MENGGUNAKAN METODE CLUSTERING DAN ASOSIASI PADA DATA TRANSAKSI PENJUALAN

Arawinda Khairunnisa, NIM.: 19106050033 (2023) DATA MINING MENGGUNAKAN METODE CLUSTERING DAN ASOSIASI PADA DATA TRANSAKSI PENJUALAN. Skripsi thesis, UIN SUNAN KALIJAGA YOGYAKARTA.

Preview	Text (DATA MINING MENGGUNAKAN METODE CLUSTERING DAN ASOSIASI PADA DATA TRANSAKSI PENJUALAN) 19106050033_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf - Published Version Download (2MB) | Preview
	Text (DATA MINING MENGGUNAKAN METODE CLUSTERING DAN ASOSIASI PADA DATA TRANSAKSI PENJUALAN) 19106050033_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf - Published Version Restricted to Registered users only Download (3MB) | Request a copy

Abstract

In this day and age, it is very easy for someone to create data or record something. However, not everyone can process the data into new information or knowledge that can be very useful. This is due to limited knowledge, tools or because they feel they do not need it. Therefore, the scale of the current data volume is very large and if the data is not processed into new information, then the data will only become garbage in memory. One way to turn data into new information is to use data mining. To overcome this problem, this research will carry out data mining using the clustering method and the association method with the K-Means and FP- Growth algorithms in processing Delima department store transaction data in February 2023 into data that can be utilized further by the store. The association method is used to find association patterns from transaction data and the clustering method is used to increase the accuracy of the association method results. The researchers chose these two algorithms because from several studies it has been proven that these two algorithms are the best algorithms for clustering and association / market basket analysis. In this study, both algorithms were successfully implemented. By using a cluster value = 2, minimum support 0.008, minimum confidence 20% and a lift > 1, it produces an association pattern and a different number of rules on data without clustering and data from clustering results. The strongest association patterns for each data are: data without clusters (Vit 600ml → Cold Drinks) with support value 0.013 and confidence value 55%, cluster0 (600ml Vit → Cold Drinks) with support value 0.010 and confidence value 50% and cluster1 (1kg Granulated Sugar, Wheat Flour → Palm Oil) with support value 0.020 and confidence value 100%. By doing data clustering before finding association patterns, it is proven to increase accuracy from 0.06887 to 28.6392.

Share this knowledge with your friends :

Actions (login required)

View Item