Hero Draft Pick Recommendation Model Based on Game Meta Using K-Nearest Neighbor Algorithm.
Abstract
Mobile Legends is a popular game that is not only enjoyed by a wide range of players but also serves as a venue for local and international e-sports competitions. One crucial factor in achieving victory is a hero draft pick strategy that aligns with the meta in a particular patch. However, many players still don't understand the importance of a balanced team composition, potentially creating teams that lack synergy and lower their chances of winning. This research aims to build a hero draft pick recommendation system using data mining techniques with the K-Nearest Neighbor (KNN) algorithm to increase the chances of winning. The dataset used consists of historical match data, including information on top picks and top hero win rates. This data is analyzed to identify optimal hero combination patterns that are effective as counter-strategies against opposing heroes. The data processing process includes data selection, preprocessing, and classification using KNN to predict potential wins based on the selected hero combinations. Test results show that the model can achieve a prediction accuracy of 75%–80%. These results demonstrate that the KNN algorithm is quite effective in identifying draft pick patterns that have the potential to increase wins, enabling the system to help players determine more optimal and balanced team compositions. The implementation of the KNN algorithm has proven effective in identifying optimal hero selection patterns, while also being a strategic solution for players to determine a more balanced team composition to significantly increase the chances of winning.
Downloads
References
U. Fayyad, G. Piatetsky-Shapiro, dan P. Smyth, “From Data Mining to Knowledge Discovery in Databases,” AI Magazine, vol. 17, no. 3, hal. 37-54, 1996.
J. Han, M. Kamber, dan J. Pei, Data Mining: Concepts and Techniques, edisi ketiga, Waltham, MA: Morgan Kaufmann, 2011.
E. Rahm dan H. H. Do, “Data cleaning: Problems and current approaches,” IEEE Data Eng. Bull., vol. 23, no. 4, hal. 3-13, 2000.
S. Patro dan K. K. Sahu, “Normalization: A Preprocessing Stage,” IARC, vol. 2, no. 3, hal. 20-22, 2015.
T. Cover dan P. Hart, “Nearest neighbor pattern classification,” IEEE Trans. Inf. Theory, vol. 13, no. 1, hal. 21–27, Jan. 1967.
T. Fawcett, “An introduction to ROC analysis,” Pattern Recognit. Lett., vol. 27, no. 8, hal. 861-874, 2006.
A. R. S. Panjaitan dan A. G. Gani, “Sistem Pakar Pemberian Rekomendasi Pemilihan Hero pada Game Mobile Legends Menggunakan Metode Forward Chaining,” Jurnal Teknologi Sistem Informasi dan Sistem Komputer Terapkan (JTEKSIS), vol. 6, no. 1, hal. 115-124, 2023.
M. R. Fachrizal dan S. Waluyo, “Penerapan Algoritma K-Nearest Neighbor untuk Klasifikasi Prediksi Kemenangan dalam E-Sports,” Jurnal Teknologi Sistem Informasi dan Sistem Komputer Terapkan (JTEKSIS), vol. 5, no. 2, hal. 210-218, 2022.
H. T. Sihotang, “Sistem Pendukung Keputusan Penentuan Ibu Rumah Tangga Penerima Bantuan Menggunakan Metode K-Nearest Neighbor,” Jurnal Teknologi Sistem Informasi dan Sistem Komputer Terapkan (JTEKSIS), vol. 4, no. 1, hal. 56-65, 2021.
Lile, F. X. K., & Suharjo, I. (2024). Optimalisasi Perbandingan Algoritma K-Nearest Neighbors dan Decision Tree untuk Prediksi Kemenangan di MPL Season 13 Mobile Legend. JATI (Jurnal Mahasiswa Teknik Informatika), 8(4), 7553-7560.
Almisbah, M. F. (2025). Sistem Prediksi Kemenangan Hero Mobile Legends Menggunakan Metode Naive Bayes (Doctoral dissertation, Universitas Lancang Kuning).
Copyright (c) 2026 Khusnul Istiqomah, Nihad Khalilullah, Rahma Oktavia, Ahmad Marsehan
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under an Attribution 4.0 International (CC BY 4.0) that allows others to share — copy and redistribute the material in any medium or format and adapt — remix, transform, and build upon the material for any purpose, even commercially with an acknowledgment of the work's authorship and initial publication in this journal.
