PREVISÃO DO CUSTO DE CARTAS EM LEGENDS OF RUNETERRA UTILIZANDO RANDOM FOREST

UMA ABORDAGEM DE APRENDIZADO DE MÁQUINA PARA ANÁLISE DE DADOS DE JOGOS ELETRÔNICOS

Authors

Keywords:

Machine learning, Random Forest, Predictive modeling, Supervised learning, Electronic games

Abstract

Artificial intelligence (AI) and machine learning (ML) are revolutionizing data analysis and decision-making in various fields, including electronic games. This study applied the Random Forest technique, a machine learning algorithm, to analyze and model card data from the game Legends of Runeterra (LoR). The main objective was to build a predictive model to estimate a card's cost based on its attributes and explore the distribution of card types in the game. The methodology involved exploratory analysis of a public dataset containing information about LoR cards, using the Python libraries Pandas, NumPy, Matplotlib, Seaborn, and Scikit-learn. The Random Forest model was trained and tested on subsets of the data, and its performance was evaluated using the mean squared error (MSE) and the coefficient of determination (R²). The results revealed insights into the distribution of cost and card types in the game, with most cards having a low cost and "Unit" type cards being the most frequent. The predictive model achieved an R² of 0.516, indicating that it explains 51.6% of the variance in card cost. This study demonstrates the potential of machine learning to analyze electronic game data and assist players and developers in making strategic decisions. The conclusions pave the way for future research exploring other machine learning algorithms and card attributes, aiming to improve the accuracy of the predictive model and deepen the understanding of game dynamics.

Published

2024-09-09

Most read articles by the same author(s)

<< < 1 2