基于迁移学习的电视观众分类模型
TV Audience Classification Model Based on Transfer Learning
卢川琴 1王国永 1张立山1
作者信息
- 1. 山东省广播电视监测中心,山东 济南 250014
- 折叠
摘要
针对智能电视机因无法获取电视观众特征数据导致无法准确地对观众进行分类的问题,提出一个基于迁移学习的电视观众分类模型.先分析用户特征和其喜好的电视节目,并训练基模型,再使用真实的电视观众数据和其喜爱观看的电视节目数据,将基类型迁移到电视观众分类任务中.实验证明,所提模型具有良好的分类准确性.
Abstract
In response to the problem that smart TV cannot obtain TV audience feature data,which makes it difficult to accurately classify viewers,this article proposes a TV audience classification model based on transfer learning. Firstly,analyze the characteristics of users and their preferred TV programs,and train a base model,then,using real TV audience data and their favorite TV programs,transfer the base type to the TV audience classification task. Experimental results have shown that the proposed model has good classification accuracy.
关键词
电视观众分类/迁移学习/深度神经网络Key words
TV audience classification/transfer learning/deep neural networks引用本文复制引用
出版年
2024
NETL
NSTL
万方数据