CAJ | 학술논문

针对数字电视信息过载问题进行了数字电视节目推荐系统的结构和推荐算法研究。构造了包含信息摄取单元、系统分析单元和推荐信息发送单元的数字电视节目推荐系统结构。提出了节目类型分析法，采用节目类型阈值和节目类型聚类两种形式，分别实现了根据经验值的人工分级和任意簇数的自动聚类，达到了有效寻找观众收视兴趣与所观看节目之间的联系的效果。提出了收视行为分析法，采用收视个体行为分析和收视群体行为分析两种形式，通过观众收视数据推测观众收视兴趣，实现了给观众个体或群体推荐感兴趣的电视节目的目的。
침대수자전시신식과재문제진행료수자전시절목추천계통적결구화추천산법연구。구조료포함신식섭취단원、계통분석단원화추천신식발송단원적수자전시절목추천계통결구。제출료절목류형분석법，채용절목류형역치화절목류형취류량충형식，분별실현료근거경험치적인공분급화임의족수적자동취류，체도료유효심조관음수시흥취여소관간절목지간적련계적효과。제출료수시행위분석법，채용수시개체행위분석화수시군체행위분석량충형식，통과관음수시수거추측관음수시흥취，실현료급관음개체혹군체추천감흥취적전시절목적목적。
To solve the difficulty in choosing TV programs caused by the information overload brought by the TV digitalization, the constitution and algorithm of a digital TV program recommendation system were studied. A structure of the system, consisting of the information importing unit, the system analyzing unit and the recommendation information sending unit, was put forward. To search the connection between audience interests and audience programmes, a programme type analytical method was proposed, which adopted the two forms of programme type threshold analyzing and programme type clustering analyzing to realize the artificial classification and automatic clustering respectively. A audience behaviour analytical method was presented. It adopted the two forms of audience behavior individuality analyzing and audience behavior group analyzing to realize the recommendation of interesting programmes for audience individualities or audience groups through inferring audience behavior interest from audience watching data.