ANALISIS KINERJA LONG SHORT TERM MEMORY UNTUK KLASIFIKASI HASIL PERTANDINGAN KLUB MANCHESTER UNITED BERDASARKAN ATRIBUT STATISTIK PERTANDINGAN

MUHAMMAD NIZHAR RIZQI MAULANA, . (2026) ANALISIS KINERJA LONG SHORT TERM MEMORY UNTUK KLASIFIKASI HASIL PERTANDINGAN KLUB MANCHESTER UNITED BERDASARKAN ATRIBUT STATISTIK PERTANDINGAN. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text COVER.pdf Download (2MB)
	Text BAB I.pdf Download (297kB)
	Text BAB II.pdf Restricted to Registered users only Download (788kB) | Request a copy
	Text BAB III.pdf Restricted to Registered users only Download (461kB) | Request a copy
	Text BAB IV.pdf Restricted to Registered users only Download (905kB) | Request a copy
	Text BAB V.pdf Restricted to Registered users only Download (334kB) | Request a copy
	Text DAFTAR PUSTAKA.pdf Download (261kB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (3MB) | Request a copy

Abstract

Pemanfaatan data statistik pertandingan sepak bola masih belum optimal dalam proses klasifikasi hasil pertandingan karena metode konvensional memiliki keterbatasan dalam menangkap pola sekuensial serta menghadapi permasalahan ketidakseimbangan data kelas (class imbalance). Kondisi ini mendorong penggunaan pendekatan deep learning seperti Long Short-Term Memory (LSTM) yang mampu memproses data runtun waktu secara lebih efektif dalam mengklasifikasikan hasil pertandingan klub Manchester United berdasarkan atribut statistik pertandingan sekaligus dibandingkan dengan metode baseline K-Nearest Neighbor (KNN). Proses penelitian dilakukan melalui tahapan Knowledge Discovery in Databases (KDD) yang meliputi data selection, preprocessing, transformation, data mining, dan evaluation, dengan menggunakan dataset pertandingan Manchester United pada English Premier League musim 2000/2001 hingga 2024/2025. Ketidakseimbangan data ditangani menggunakan teknik Synthetic Minority Over-sampling Technique (SMOTE), sedangkan evaluasi model dilakukan menggunakan confusion matrix dengan metrik akurasi, precision, recall, dan F1-score. Hasil penelitian menunjukkan bahwa sebelum penerapan SMOTE, model LSTM mengalami overfitting dengan akurasi pengujian sekitar 70%, sementara KNN memiliki performa yang lebih rendah. Setelah penerapan SMOTE, performa kedua model meningkat secara signifikan, di mana LSTM mencapai akurasi sekitar 85% dan menunjukkan kestabilan yang lebih baik dibandingkan KNN. Temuan ini menunjukkan bahwa LSTM lebih efektif dalam menangani data sekuensial dan, ketika dikombinasikan dengan SMOTE, mampu meningkatkan kemampuan generalisasi model secara signifikan, sehingga berpotensi menjadi pendekatan yang lebih akurat dalam klasifikasi hasil pertandingan sepak bola. Kata kunci: Klasifikasi, KNN, LSTM, SMOTE ***** The utilization of statistical data in football matches has not been fully optimized for match outcome classification, as conventional methods still struggle to capture sequential patterns and often face class imbalance issues. This condition encourages the use of deep learning approaches such as Long Short-Term Memory (LSTM), which are capable of processing time-series data more effectively to classify match outcomes of Manchester United based on statistical attributes, while also being compared with the baseline method K-Nearest Neighbor (KNN). The research process is conducted through the Knowledge Discovery in Databases (KDD) stages, including data selection, preprocessing, transformation, data mining, and evaluation, using historical match data from the English Premier League spanning the 2000/2001 to 2024/2025 seasons. Data imbalance is addressed using the Synthetic Minority Over-sampling Technique (SMOTE), while model performance is evaluated using a confusion matrix with accuracy, precision, recall, and F1-score metrics. The results indicate that prior to applying SMOTE, the LSTM model experienced overfitting with testing accuracy of approximately 70%, whereas KNN showed lower performance. After applying SMOTE, both models demonstrated significant improvement, with LSTM achieving approximately 85% accuracy and showing more stable performance compared to KNN. These findings indicate that LSTM is more effective in handling sequential data and, when combined with SMOTE, significantly improves the model’s generalization ability, making it a more reliable approach for football match outcome classification. Keywords: Classification, KNN, LSTM, SMOTE

Item Type:	Thesis (Sarjana)
Additional Information:	1). Dr. Widodo, M.Kom. ; 2). Neng Ayu Herawati, S.Pd., M.T.
Subjects:	Sains > Matematika > Ilmu Komputer Teknologi dan Ilmu Terapan > Teknik Komputer
Divisions:	FT > S1 Pendidikan Teknik Informatika Komputer
Depositing User:	Muhammad Nizhar Rizqi Maulana .
Date Deposited:	24 Jul 2026 04:09
Last Modified:	24 Jul 2026 04:09
URI:	http://repository.unj.ac.id/id/eprint/67104

Actions (login required)

View Item