一种基于实时CTR的移动应用商店内容推荐改进算法  

An Improvement of Content-Recommend Algorithm Based on Real-time CTR in Mobile Market

作　　者：冯欣[1] 夏旸[1] FENG Xin XIA Yang(School of Computer Science and Technology, Changchun University of Science and Technology, Changchun 130022)

机构地区：[1]长春理工大学计算机科学技术学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2017年第2期122-126,共5页Journal of Changchun University of Science and Technology(Natural Science Edition)

摘　　要：针对内容信息过载,冷启动等导致移动应用市场用户消费受限、广告收入受阻的问题,文章提供一种能够提高移动应用市场人均分发能力的内容推荐算法。首先,收集一段时间内产生的内容推荐数据,作为待处理的推荐内容集合。然后,通过一种改进的实时CTR推荐算法,对已有内容进行基于展示、点击、下载的重新排列,并将重新排列的数据展示在移动应用市场内部。与传统的CTR推荐算法相比较,改进后的实时CTR推荐算法在评价维度上更加合理。通过对比,改进后的实时CTR推荐算法可以提高移动应用市场的分发能力,适用于信息过载下的移动应用市场。For the content and information overload,cold start and others as results of the limitation of mobile application market users＇ consumption and the obstruction of advertise revenue,in this paper,a content-recommend algorithm to improve the consumption ability of the mobile application market for each consumer is provided. First,the recommended content datum generated within the period are collected as the pending set. Then,through an improved real-time CTR recommendation algorithm,the existing contents based on their impressions are rearranged,clicked and downloaded,then the result in mobile application market is displayed. Compared with the traditional CTR recommendation algorithm,the improved real-time CTR recommendation algorithm is more reasonable in the evaluation dimensions.By contrast,the improved real-time recommendation algorithm can improve the distribution capabilities of the mobile application market,especially for those with the problem of information overload.

关 键 词：移动应用市场 内容推荐 CTR 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]