Skripsi

KLASIFIKASI TALENT SCOUTING RENANG MENGGUNAKAN HYBRID BACKPROPAGATION DAN ALGORITMA GENETIKA

---

Conventional talent scouting systems for potential swimming athletes still rely on subjective assessments by coaches, potentially leading to bias and measurement errors. To address this issue, this study developed a Hybrid Backpropagation–Genetic Algorithm-based classification system to identify potential athletes based on distance specialization. The data used consisted of 100 potential athletes from the Prabumulih City PSC and the Palembang City TSAC, with 13 initial attributes processed into 12 relevant attributes. The potential label classification was divided into two categories: “JJ” (long distance) and “JP” (short distance). Backpropagation is good at classification but has limitations such as slow convergence and the risk of getting stuck in a local minimum, especially on high-dimensional data with irrelevant features, thus failing to achieve a global optimal solution. Therefore, a Genetic Algorithm was applied to select optimal features using a binary chromosome representation, while Backpropagation was used as a fitness evaluator based on validation accuracy. Although the hybrid approach requires longer computational time due to evaluation at each generation, this method has been proven to improve accuracy and reduce overfitting. Parameter testing was conducted using the One Factor at a Time (OFAT) method and the stability of the results was maintained by calculating the average of the five best performances. Standard backpropagation produced 82% accuracy with a learning rate of 0.09, 400 epochs, and 20 hidden units, while the Hybrid model achieved 92% with a population of 40, 10 generations, a crossover rate of 0.5, and a mutation rate of 0.02 with a precision of 84%, a recall of 100%, and an F1-score of 91.11%. The 10% increase in accuracy demonstrates the effectiveness of the Genetic Algorithm in improving the performance of swimming talent scouting classification.

---

Availability

Detail Information

Series Title

-

Call Number

T1950762026

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

xxxv, V-15, 163 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

796.015 07

Content Type

Text

Media Type

unmediated

Carrier Type

-

Edition

-

Subject(s)

Prodi Teknik Informatika
Pencarian Bakat (Talent Scouting)

Specific Detail Info

-

Statement of Responsibility

KA

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment