Comparison of K-Medoids Algorithm with K-Means on Number of Student Dropped Out

Aziz, Muhammad Abdul; Hidayat, Yusuf Amin; Febrianti, Diah Ratna; Aida, Amrina Nur; Amalia, Lisa; Tahyudin, Imam; Darmayanti, Irma

doi:10.1109/apics56469.2022.9918789

Cited by 1 publication

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tables 7 and 8 show the performance of the optimal number of clusters using the DBI and Elbow methods [32]- [34]. In general, this research proves that the kmeans method can handle large amounts of data quickly, effectively, efficiently, and with very optimal modeling accuracy [21]- [24]. The limitation of this technique is determining the number of clusters in big data, where additional methods like DBI or Elbow must be used to obtain the optimal number of clusters.…”

Section: Figure 12 Visualization Of Model Clustering Results With K=3mentioning

confidence: 96%

“…This was done because the k-means approach can handle large amounts of data quickly and effectively [21]. This technique is very efficient because the computational cost required is not too high compared to Gaussian Mixture [22], In addition, k-means can produce optimal accuracy compared to the k-medoids method, this is according to research on clustering student demographic data [23] and clustering students who drop out of school [24]. Therefore, in this study, data mining with the k-means technique is carried out based on sales data of automotive products in Indonesia obtained from the Gaikindo organization website.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Data Mining Modeling in Clustering Car Products Sales Data in the Automotive Industry in Indonesia

Astuti,

Yuniarti

2023

j. manaj. ind. dan logist.

View full text Add to dashboard Cite

The research aims to build a model based on sales data for all automotive products in Indonesia using data mining with a k-means approach. This study uses automotive product sales data from January 2017 to September 2022. The lowest Davis-Bouldin index shows that three clusters (k=3) have the best performance. Based on the clustering results, 92% of the items are in cluster 0, 1% in cluster 1, and 7% in cluster 2. In addition, the clustering results show that cluster 1 is a car product with high sales volume. Cluster 2 is a car product with medium sales volume. Furthermore, cluster 0 is a car product with low sales volume. Business people or related parties can use data visualization and extraction from clustering results to learn the latest insights and information in determining business strategies, policies, and decisions to improve business competitiveness.

show abstract

Section: Figure 12 Visualization Of Model Clustering Results With K=3mentioning

confidence: 96%