论文部分内容阅读
随着电视节目日益丰富,电视用户正面临与互联网用户类似的“信息过载”问题,如何帮助用户及时收看到所需的节目?节目推荐系统就可很好地解决这一问题。基于真实的商用双向有线电视网络平台建立的集用户收视行为数据采集、用户收视特征分析、节目推荐、节目搜索、节目预定和节目收视率分析于一体的节目推荐系统,能及时跟踪用户收视兴趣变化并对所推荐的节目进行自动调整。本项目所研究发明的用户特征建模、节目特征建模技术、用户特征提取算法及节目推荐算法的技术水平在国内外处于领先地位。系统所采集的用户收视行为数据可真实反映用户的消费倾向,未来可结合BOSS系统的客户资料及计费数据,利用数据挖掘技术,为CRM系统、经营分析系统等提供基础数据,从而实现精确营销。
With the increasingly rich television programs, television users are faced with the problem of “information overload ” similar to that of Internet users. How to help users to watch the desired programs in time? The program recommendation system can solve this problem well. Based on the real commercial two-way cable network platform, this system integrates user’s viewing behavior data collection, user’s viewing characteristics analysis, program recommendation, program search, program reservation and program rating analysis in one program recommendation system, And the recommended programs automatically adjust. The research on this project features user modeling, program feature modeling, user feature extraction algorithm and program recommendation algorithm in a leading position both at home and abroad. The data collected by the system can truly reflect the user’s propensity to consume. In the future, the data can be combined with the customer data and billing data of the BOSS system to provide basic data for the CRM system and the business analysis system by means of data mining technology so as to achieve accurate marketing .