Pembentukan Target Pasar Berdasarkan Data Stream Transaksi Kartu Kredit (Clustering dan Association Rule) pada PT Bank Bukopin

Abstract

the purpose of research is to analyze the formation of the target market customer segmentation based on job characteristics, income, education, age, region of origin, and patterns of credit card merchant. The data were analyzed using two data mining techniques of clustering with K-Means and Association Rule Mining (ARM) and supported by Apriori and Random Sampling technique with the Slovin formula and Principal Component Analysis (PCA). The clustering tests in 10 replications on the sampling of 350 clients supported by Pca produced the three best clusters that had the silhouette value close to 1 i.e. 0.39 to 0.40. meanwhile, the arm testing with apriori using a minimum support of 1% and a minimum confidence of 40% produced two patterns of credit card merchant transactions. In the first pattern, when the Hotel merchant type (hhl) was transacted, the Restaurant merchant type (RRT) was also transacted, and in the second pattern, if the Service Station merchant type (RSS) was transacted, the Restaurant merchant type (RRT) was also transacted. The three clusters and two types of merchant patterns obtained can generate inputs for the company to identify its potential customers based on the characteristics of the target customers by connecting them to the merchant type pattern frequently used.

K-Means Clustering Method for Determining Waste Transportation Routes to Landfill

“…This stage includes identifying missing values in the data to find out if there is incomplete data, selecting data to determine which data will be used in the clustering process, deleting irrelevant data, and then performing the data normalization stage. Data normalization plays an essential role in the data preprocessing stage because it can reduce the difference in scale values between variables which can cause some variables to be too dominant or disproportionate to the analysis results (Riza et al, 2018). In addition, data normalization can also improve the accuracy of a cluster.…”

Customer Segmentation For Digital Marketing Based on Shopping Patterns

“…In the marketing world, every customer is unique because they have different behaviors and preferences. Therefore, customer segmentation is needed as a key marketing strategy (Riza, Seminar and Maulana, 2018). Customer segmentation can help marketers to find correlations between certain customer segments with their shopping patterns and preferences so that marketers can adjust strategies, messages, promotions and special offers for these customer segments (Tabianan, Velu and Ravi, 2022).…”

“…Machine learning enables marketers to analyze customer shopping history data in real-time, follow patterns and dynamics of changes in customer preferences (Aryuni, Madyatmadja and Miranda, 2018;Tabianan, Velu and Ravi, 2022;Sun, Liu and Gao, 2023). Riza, et al used the k-means and apriori methods to segment potential customers based on customer characteristics and transaction patterns with certain types of merchants (Riza, Seminar and Maulana, 2018). Lee, et al applied several machine learning approaches such as k-mean, fuzzy c-mean, and Wald's method for customer segmentation (Lee et al 2021).…”