Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Jurnal Teknik Informatika (JUTIF)
Lombu, Anas Satria
Unknown Affiliation
Computer Science & IT

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
PREDICTING FANTASY PREMIER LEAGUE POINTS USING CONVOLUTIONAL NEURAL NETWORK AND LONG SHORT TERM MEMORY Lombu, Anas Satria; Paputungan, Irving Vitra; Dewa, Chandra Kusuma
Jurnal Teknik Informatika (Jutif) Vol. 5 No. 1 (2024): JUTIF Volume 5, Number 1, February 2024
Publisher : Informatika, Universitas Jenderal Soedirman

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52436/1.jutif.2024.5.1.1792

Abstract

Fantasy Premier League is a fantasy sports-based game focused on football, particularly the English Premier League. each manager in this game is given the opportunity to build a virtual team for one season. A virtual team consists of various player positions that will earn FPL points based on their real-word performance. This research aims to implement a deep learning algorithm to predict FPL points generated by players based on their performance in the last 5 matches using a dataset collected from August 14, 2021, to May 21, 2022. The prediction model is designed using a Convolutional Neural Network algorithm consisting of one-dimensional Convolution layers, Max Pooling, and Dense layers. Additionally, a Long Short Term Memory algorithm with LSTM layers and Dense layers totaling 64 units is added as a comparison model to determine the best performing deep learning model in this study. In the first scenario, with a 70:30 data ratio, the average Mean Squared Error values obtained for 4 player positions using CNN are 0.0052 and 0.0027 for LSTM. Meanwhile, in the second scenario with an 80:20 data ratio, the evaluation results are 0.0027 for CNN and 0.0022 for LSTM. the model evaluation results indicate that the LSTM algorithm, utilizing three gates in the model architecture, is superior in recognizing historical data sequences compared to the CNN algorithm.
Co-Authors Dewa, Chandra Kusuma Paputungan, Irving Vitra