CAJ | 학술논문

为更好的让智能终端的新闻阅读有针对性，本课题对新闻个性化推荐技术展开了研究，设计并实现了基于 Android 智能终端的个性化新闻推荐系统。系统采用 C ／S 架构，服务器端首先进行新闻爬取和分类建模，然后根据新闻相似性和用户阅读行为，分阶段利用三种推荐模型即基于新闻分类的粗粒度推荐、基于内容的新闻推荐及基于内容和基于用户协同过滤混合的新闻推荐进行个性化新闻推荐；客户端负责个性化新闻阅读以及用户行为数据上传。经测试，本系统的分阶段不同个性化推荐算法可以很好的满足广大用户的个性化新闻阅读需求。
위경호적양지능종단적신문열독유침대성，본과제대신문개성화추천기술전개료연구，설계병실현료기우 Android 지능종단적개성화신문추천계통。계통채용 C ／S 가구，복무기단수선진행신문파취화분류건모，연후근거신문상사성화용호열독행위，분계단이용삼충추천모형즉기우신문분류적조립도추천、기우내용적신문추천급기우내용화기우용호협동과려혼합적신문추천진행개성화신문추천；객호단부책개성화신문열독이급용호행위수거상전。경측시，본계통적분계단불동개성화추천산법가이흔호적만족엄대용호적개성화신문열독수구。
In order to make the news reading in the intelligent terminal more personalized,we design and implement the system of personalized news recommendation based on the Android intelligent terminal.The system,based on C /S structure,firstly crawls news and establishes the models with classifications in the server terminal,and then,in different phases to carry out the personalized news recommendation in accordance with similarities between news and users'reading behaviors,uses different recommendation models such as the coarse -grained recommendation based on news classification,the news recommendation based on content and the mixed recommendation based on news content and users collaborative filtering.The client terminal is used for reading personalized news and uploading user behavior data.After test,the personalized recommendation algorithm of the system in different phases can satisfy the users'reading needs for the personalized news.