Sales prediction using machine learning approaches

Saraswathi, K.; Renukadevi, N. T.; Nandhinidevi, S.; Gayathridevi, S.; Naveen, P.

doi:10.1063/5.0068655

Cited by 5 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Or perhaps many previous studies only used traditional neural networks and regression analysis to predict the sales of electronic games. Although XGboost and LightGBM gradient boosting tree algorithms have been widely used in predicting many economic problems such as stock prices and housing prices, few people have previously used these two methods to analyze and predict sales in the gaming market [3,4].…”

Section: Introductionmentioning

confidence: 99%

Research on the Prediction on the Sales of Electronic Games

Huang

2023

HBEM

View full text Add to dashboard Cite

This essay's primary goal is to predict and analyze the sales characteristics of electronic games based on different features, including but not limited to release time, ranking, publisher, and game genre. The methods employed in this paper include a neural network architecture where the activation function is ReLU , XGBoost model, and LightGBM model. Through experiments, this paper effectively constructs machine learning models using these three methods to predict video game sales. Furthermore, the performance of these three methods is compared and analyzed using three evaluation metrics: root mean squared error (RMSE), mean absolute error (MAE), and mean squared error (MSE). And the prediction error of each model is plotted as a related line chart to visually show the details of the three methods in analyzing this problem. Finally, by comparing the experimental results of these three methods in the field of electronic game sales prediction, this article analyzes the advantages and disadvantages of neural networks, XGBoost, and LightGBM in related problem analysis, providing readers with reference and reference.

show abstract

Section: Introductionmentioning

confidence: 99%