Analisis Kecocokan Antar Pemain Sepakbola Menggunakan Machine Learning.

Winata, Patrick Cipta (2022) Analisis Kecocokan Antar Pemain Sepakbola Menggunakan Machine Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05111840000098-Undergraduate_thesis.pdf Restricted to Repository staff only Download (3MB)

Abstract

Sistem rekrutmen pemain oleh klub sepakbola hingga saat ini dapat dilihat masih berdasarkan pada kualitas individual dan performa pemain saja. Staff yang ditugaskan untuk memantau pemain untuk direkrut belum mempertimbangkan kecocokan pemain tersebut dengan tim yang menginginkannya. Disini, dibutuhkan suatu pertimbangan untuk menilai seberapa potensi kecocokan antar pemain. Namun, untuk menilai potensi kecocokan pemain dengan klub yang dituju susah karena alat analitik yang ada di bidang sepakbola masih hanya berfokus ke evaluasi pemain secara individual dan tidak mengarah ke interaksi antar pemain. Dari masalah tersebut, dicoba untuk mengetahui seberapa baik seorang pemain bermain dalam suatu tim dengan memprediksi 2 metrics, yaitu Joint Offensive Impact (JOI) dan Joint Defensive Impact (JDI). Kecocokan antar pemain dalam suatu tim akan dinilai dan dapat diketahui pemain mana yang cocok untuk bermain bersama berdasarkan nilai JOI dan JDI nya. Nilai JOI digunakan sebagai tanda kecocokan kedua pemain dalam menyerang, sedangkan JDI digunakan sebagai tanda kecocokan kedua pemain dalam bertahan. Setelah itu, dapat dimasukkan case untuk rekrutmen pemain dimana menilai kecocokan pemain yang akan direkrut kedalam tim. Tahapan yang akan dilakukan diantaranya pemrosesan data dari data event sepakbola menjadi bentuk yang lebih relevan. Setelah data diproses, dilakukan perhitungan scoring metrics yang akan dibutuhkan untuk menentukan kecocokan antar pemain. Data yang telah melewati proses scoring akan digunakan untuk proses training model dan model yang telah di train akan digunakan untuk memprediksi nilai kecocokan antar pemain. Akan dilakukan juga evaluasi pada hasil yang telah diprediksi model untuk pengecekan performa model. Hasil yang diperoleh dari implementasi sistem ini adalah model regresi yang mampu memprediksi nilai metrics JOI serta model yang mampu memprediksi nilai metrics JDI. Kedua model kemudian akan dievaluasi menggunakan metrics Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), dan R Squared. Model terbaik memiliki nilai RMSE terendah terhadarp test set yaitu 0,0449. Model ini merupakan model Catboost yang telah melalui proses parameter tuning menggunakan Grid Search.
==============================================================================================================================
Football Club’s player recruitment system so far can be seen that it is still based on individual quality and player performance. The staff assigned to monitor players for recruitment have not considered the compatibility of these players with the team that wants them. Here, it takes a consideration to assess how much potential matches between players. However, assessing a player's potential compatibility with the target club is difficult the existing analytical tools in the field of football still only focus on evaluating individual players and do not lead to interactions between players. From this problem, we try to find out how well a player plays in a team by predicting 2 metrics, namely Joint Offensive Impact (JOI) and Joint Defensive Impact (JDI). The compatibility between players in a team will be assessed and it can be seen which players are suitable to play together based on their JOI and JDI values. The JOI value is used as a sign of the compatibility of the two players in attack, while the JDI is used as a sign of the compatibility of the two players in defense. After that, a case for player recruitment can be used which assesses the compatibility of the player to be recruited into the team. The process that will be carried out include processing data from football event data into a more relevant form. After the data is processed, scoring metrics are calculated which will be needed to determine the match between players. The data that has passed the scoring process will be used for the model training process and the model that has been trained will be used to predict the match value between players. There will also be an evaluation of the results that have been predicted by the model to check the model's performance. The results obtained from the implementation of this system are a regression model that can predict the value of the JOI metrics and a model that is able to predict the value of the JDI metrics. Both models will then be evaluated using the metrics Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R Squared. The best model has the lowest RMSE value against the test set, which is 0,0449. This model is a Catboost model that has gone through the parameter tuning process using Grid Search.

Item Type:	Thesis (Other)
Additional Information:	RSIf 006.31 Win a-1 2022
Uncontrolled Keywords:	Sepakbola, JOI, JDI, Kecocokan, Catboost. Football, JOI, JDI, Compatibility,Catboost.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	26 May 2026 02:29
Last Modified:	26 May 2026 02:29
URI:	http://repository.its.ac.id/id/eprint/133422

Actions (login required)

View Item