Sinadia, Herbie (2022) Sports analytics: profiling and movement analysis for badminton athletes = Sports analytics: profiling dan analisa pergerakan untuk atlet bulutangkis. Masters thesis, Universitas Pelita Harapan.
|
Text (Title)
Title.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (65kB) | Preview |
|
|
Text (Abstract)
Abstract.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (255kB) | Preview |
|
|
Text (ToC)
ToC.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (620kB) | Preview |
|
|
Text (Chapter1)
Chapter1.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (797kB) | Preview |
|
![]() |
Text (Chapter2)
Chapter2.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (1MB) |
|
![]() |
Text (Chapter3)
Chapter3.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (2MB) |
|
![]() |
Text (Chapter4)
Chapter4.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (571kB) |
|
![]() |
Text (Chapter5)
Chapter5.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (258kB) |
|
|
Text (Bibliography)
Bibliography.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (569kB) | Preview |
|
![]() |
Text (Appendices)
Binder1_watermark.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (6MB) |
Abstract
Sports analytics have proven be to be an important tool for sports players and practitioners to remain competitive in their respective fields. In sports analytics, we can utilize artificial intelligence to assess an athlete or sports team’s performances in detail. Badminton is a field of sports where an athlete can greatly benefit from sports analytics. However, we observed that there is not enough research on badminton sports analytics, especially when compared to other popular sports such as soccer and basketball. Through this research, we proposed a method to perform profiling on badminton athletes by utilizing artificial intelligence to analyze a player’s performance from historical match data (score analysis) and computer vision to detect and track certain players in badminton match videos (video analysis). We used SVM, Random Forest, K-means, and hierarchical clustering to perform regression, classification, and clustering on players’ match dataset. For the computer vision part, we trained a custom YOLOv5 model to detect players, serving positions, and shuttlecocks on a badminton match video and then using Tracker to track and calculate the player and shuttlecock’s movements. Results shows that we have successfully built a functioning model of badminton athlete profiling. For score analysis, Random Forest and SVM with the Linear kernel are the best performing models on regression and classification, with an average R2 regression score of 0.9705 and 0.9109 and an average classification accuracy of 96.53% and 96.64% across all test cases for both models respectively. K-means and hierarchical clustering performed very similarly, thus both models can be used to effectively perform clustering on an athlete’s match results. For video analysis, our custom-trained YOLOv5 model was able to detect players, shuttlecocks, and serve positions in most of the video’s frames, and object tracking with Tracker was proven to be effective in helping us understand a badminton athlete’s performances. / Sports analytics telah terbukti menjadi metode yang efektif agar atlet olahraga dapat tetap bersaing di bidangnya masing-masing. Dalam sports analytics, kecerdasan buatan dapat digunakan untuk mempelajari kemampuan dan performa suatu atlet atau tim olahraga dengan detil. Salah satu cabang olahraga dimana sports analytics dapat membantu para atletnya adalah bulutangkis. Namun, kami melihat bahwa penelitian mengenai sports analytics dalam bulutangkis masih belum banyak, terutama ketika dibandingkan dengan cabang olahraga popular lainnya seperti sepak bola dan basket. Dalam penelitian ini, kami mengusulkan suatu metode untuk melakukan profiling pada atlet bulutangkis menggunakan kecerdasan buatan untuk analisis kemampuan seorang pemain dengan melihat hasil pertandingan historis (score analysis) dan visi komputer untuk mendeteksi dan melacak pergerakan pemain dalam video pertandingan bulutangkis (video analysis). Untuk bagian score analysis, kami menggunakan SVM, Random Forest, K-means, dan hierarchical clustering untuk melakukan regression, classification, dan clustering dari dataset hasil pertandingan pemain. Untuk bagian video analysis, kami melatih suatu model YOLOv5 untuk mendeteksi pemain, shuttlecock, dan posisi serve pada video pertandingan bulutangkis dan dilanjutkan dengan penggunaan Tracker untuk melacak dan menghitung pergerakan atlet dan shuttlecock. Hasil menunjukkan bahwa kami telah berhasil mengembangkan model untuk melakukan profiling terhadap atlet bulutangkis. Untuk score analysis¸ Random Forest dan SVM merupakan model-model yang memiliki performa terkuat dalam regression dan classification, dengan rata-rata skor adjusted R2 sebesar 0.9705 and 0.9109 dan rata-rata akurasi klasifikasi sebesar 96.53% and 96.64% dari semua uji coba untuk kedua model tersebut. Untuk video analysis, model YOLOv5 yang telah kami latih dapat mendeteksi pemain, shuttlecock, dan posisi serve pada sebagian besar frame video pertandingan bulutangkis yang digunakan, dan pelacakkan pergerakan obyek dengan Tracker terbukti efektif dalam membantu kami memahami performa suatu atlet bulutangkis.
Item Type: | Thesis (Masters) | ||||||||
---|---|---|---|---|---|---|---|---|---|
Creators: |
|
||||||||
Contributors: |
|
||||||||
Uncontrolled Keywords: | sports analytics ; badminton ; regression ; classification ; clustering ; computer vision ; yolov5 ; tracker | ||||||||
Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science | ||||||||
Divisions: | University Subject > Current > Faculty/School - UPH Karawaci > School of Information Science and Technology > Master of Informatics Current > Faculty/School - UPH Karawaci > School of Information Science and Technology > Master of Informatics |
||||||||
Depositing User: | Users 27818 not found. | ||||||||
Date Deposited: | 09 Feb 2023 06:25 | ||||||||
Last Modified: | 14 Feb 2023 09:28 | ||||||||
URI: | http://repository.uph.edu/id/eprint/54144 |
Actions (login required)
![]() |
View Item |