摘要
随着网络游戏的快速兴起,精确的游戏销量预测具有较高的商业价值,能够明确各方投资方向,提高收益,形成合作共赢。本文以影响游戏销量的特征数据为样本,建立基于梯度提升决策树(Gradient Boosting Decision Tree,GBDT)算法的游戏销量预测模型;并将GBDT模型预测结果与决策树、线性回归、极端随机树进行对比分析。分析表明,本文所建立的游戏销量预测模型较其它预测模型具有较高的拟合优度,预测效果更好,且在预测阶段的计算速度快,在分布稠密的数据集上,泛化能力和表达能力较好。
With the rapid rise of online games,accurate prediction of game sales has high commercial value,which can clarify the investment direction of all parties,improve revenue and form win-win cooperation.Therefore,this paper takes the feature data that affect game sales as the sample and establishes a game sales prediction model based on the Gradient Boosting Decision Tree(GBDT)algorithm.The prediction results of GBDT model are compared with Decision Tree,Linear Regression and ExtraTree.The final results show that compared with other prediction models,the established game sales prediction model has higher goodness of fit,better prediction effect,faster calculation speed in the prediction stage,and better generalization and expression ability in dense distribution data sets.
作者
徐英卓
郭博
王六鹏
XU Yingzhuo;GUO Bo;WANG Liupeng(School of Computing,Xi′an Shiyou University,Xi′an 710000,China;School of Petroleum Engineering,Xi′an Shiyou University,Xi′an 710000,China)
出处
《智能计算机与应用》
2023年第1期182-185,共4页
Intelligent Computer and Applications
关键词
游戏销量
预测
梯度提升决策树
game sales
forecast
Gradient lifting decision tree