DEWIMURNIATI, PUTRI AYU (2025) PREDIKSI KEMENANGAN TIM DALAM TURNAMEN VALORANT CHAMPIONS TOUR 2024 BERDASARKAN STATISTIK PERFORMA PEMAIN. S1 thesis, Universitas Mercu Buana Jakarta - Menteng.
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (COVER)
41521010050-Putri Ayu Dewimurniati-01 Cover - Putri Ayu.pdf Download (450kB) |
|
Text (BAB I)
41521010050-Putri Ayu Dewimurniati-02 BAB 1 - Putri Ayu.pdf Restricted to Registered users only Download (49kB) |
|
Text (BAB II)
41521010050-Putri Ayu Dewimurniati-03 BAB 2 - Putri Ayu.pdf Restricted to Registered users only Download (269kB) |
|
Text (BAB III)
41521010050-Putri Ayu Dewimurniati-04 BAB 3 - Putri Ayu.pdf Restricted to Registered users only Download (134kB) |
|
Text (BAB IV)
41521010050-Putri Ayu Dewimurniati-05 BAB 4 - Putri Ayu.pdf Restricted to Registered users only Download (308kB) |
|
Text (BAB V)
41521010050-Putri Ayu Dewimurniati-06 BAB 5 - Putri Ayu.pdf Restricted to Registered users only Download (106kB) |
|
Text (DAFTAR PUSTAKA)
41521010050-Putri Ayu Dewimurniati-08 Daftar Pustaka - Putri Ayu.pdf Restricted to Registered users only Download (104kB) |
|
Text (LAMPIRAN)
41521010050-Putri Ayu Dewimurniati-09 Lampiran - Putri Ayu.pdf Restricted to Registered users only Download (1MB) |
Abstract
Dalam beberapa tahun terakhir, industri E-sports berkembang pesat, dengan turnamen besar seperti Valorant Champions Tour (VCT) menarik perhatian global. Penelitian ini bertujuan membangun model prediksi kemenangan tim berdasarkan performa pemain VCT 2024 menggunakan algoritma Support Vector Machine (SVM) dan Random Forest. Evaluasi dilakukan melalui akurasi, presisi, recall, AUC, serta visualisasi interpretasi model. Untuk SVM dengan kernel RBF menunjukkan akurasi 82–85% dan AUC 0,85– 0,90, menandakan kemampuan klasifikasi yang tinggi, meski interpretasi fitur terbatas karena sifat non-linear. Sementara itu, Random Forest mencatat akurasi serupa (80–85%) dan unggul dalam interpretasi fitur, dengan ACS, rasio K:D, dan KAST sebagai prediktor paling signifikan.Validasi menggunakan GroupKFold menghasilkan akurasi rata-rata 83% dengan deviasi sekitar 2% dan selisih akurasi train-test di bawah 5%, menunjukkan model stabil dan tidak overfit. Temuan ini menegaskan bahwa agresivitas individu dan kerja sama tim (KAST) adalah faktor utama kemenangan. Secara praktis, model ini berguna bagi pelatih untuk fokus pada ACS dan koordinasi tim. Random Forest cocok untuk analisis strategis mendalam, sedangkan SVM efektif untuk prediksi cepat. Kedepan, kombinasi keduanya melalui ensemble model berpotensi meningkatkan akurasi di atas 90%. Penelitian ini juga memberikan dasar analisis yang dapat diterapkan pada game kompetitif lainnya serta membuka peluang untuk riset lanjutan seperti prediksi kemenangan real-time atau analisis performa berdasarkan peran pemain. In recent years, the E-sports industry has grown rapidly, with major tournaments like the Valorant Champions Tour (VCT) gaining global attention. This study aims to build a team win prediction model based on player performance data from the VCT 2024 using Support Vector Machine (SVM) and Random Forest algorithms. The models were evaluated using accuracy, precision, recall, AUC, and visual interpretation. For SVM with an RBF kernel achieved an accuracy of 82–85% and an AUC of 0.85–0.90, indicating strong classification performance, although feature interpretation is limited due to the non-linear nature of the kernel. Meanwhile, the Random Forest model achieved similar accuracy (80–85%) and excelled in feature interpretability, identifying ADR, K\:D ratio, and KAST as the most significant predictors. Validation using GroupKFold resulted in an average accuracy of 83% with a standard deviation of around 2%, and a train-test accuracy gap of less than 5%, indicating a stable model with low overfitting risk. The findings confirm that individual aggressiveness and team coordination (KAST) are key factors in determining victory. Practically, these models can guide coaches to focus training on improving ACS and team synergy. Random Forest is suitable for in-depth strategic analysis, while SVM is more effective for fast predictions with limited data. In the future, combining both into an ensemble model could increase accuracy beyond 90%. This study also offers a transferable analytical framework for other competitive games and opens opportunities for further research, such as real-time win prediction or player-specific performance analysis
| Item Type: | Thesis (S1) |
|---|---|
| NIM/NIDN Creators: | 41521010050 |
| Uncontrolled Keywords: | Prediksi Kemenangan, Valorant Champions Tour 2024, Statistik Performa Pemain, Machine Learning, E-sports. Prediction of Victory, Valorant Champions Tour 2024, Player Performance Statistics, Machine Learning, E-sports. |
| Subjects: | 000 Computer Science, Information and General Works/Ilmu Komputer, Informasi, dan Karya Umum > 000. Computer Science, Information and General Works/Ilmu Komputer, Informasi, dan Karya Umum > 004 Data Processing, Computer Science/Pemrosesan Data, Ilmu Komputer, Teknik Informatika |
| Divisions: | Fakultas Ilmu Komputer > Informatika |
| Depositing User: | NAIMAH NUR ISLAMIDIYANAH |
| Date Deposited: | 15 Sep 2025 03:52 |
| Last Modified: | 15 Sep 2025 03:52 |
| URI: | http://repository.mercubuana.ac.id/id/eprint/97872 |
Actions (login required)
 |
View Item |
