基于深度神经网络的影视评分预测  

Prediction of film ratings based on Deep Neural Network

作　　者：朱衍熹 张明西[1] 赵瑞 许星波 ZHU Yanxi;ZHANG Mingxi;ZHAO Rui;XU Xingbo(College of Communication and Art Design,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学出版印刷与艺术设计学院,上海200093

出　　处：《智能计算机与应用》2024年第6期79-87,共9页Intelligent Computer and Applications

基　　金：国家重点研发计划项目(2021YFF0900400);国家自然科学基金(62002225);上海市自然科学基金(21ZR1445400)。

摘　　要：影视评分能直接反映影视作品的上映效果或收益情况,然而目前影视特征的提取方法单一,信息挖掘不充分。针对这一问题,提出一种基于混合特征表示向量的深度神经网络影视评分预测模型。根据影视作品的属性特征通过词袋模型、特征拆分、TF-IDF文本矢量化方法生成影视混合特征表示向量,并构建基于深度神经网络的影视评分预测模型。实验结果表明:测试集MAE、MSE、SmoothL1 Loss指标值在模型100轮迭代训练后收敛,MAE为0.82,MSE为1.07,SmoothL1Loss为0.45,验证了所提方法对影视作品的评分预测有很好效果,能有效评估影视作品上映后的价值。Film ratings can directly reflect the screening effect or income of film.However,the current method of extracting film features is single and the information mining is insufficient.To solve this problem,a film ratings prediction model based on Deep Neural Network and mixed feature representation vector is proposed.According to the attributes of the film,the representation vectors of film mixed features are generated by bag of words model,feature splitting and TF-IDF,and the film ratings prediction model based on Deep Neural Network is constructed by using vectors.The experimental results show that the index values of MAE,MSE and SmoothL1 Loss convergence after 100 iterations of training,MAE is 0.82,MSE is 1.07,SmoothL1 Loss is 0.45,which proves that the proposed method has a good effect on the ratings prediction of the film,and can effectively evaluate the value of the film after release.

关 键 词：深度神经网络 词袋模型 TF-IDF 影视评分 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]