基于主题词频共现的渔业科学数据推荐  

Fishery Science Data Recommendation Based on Frequency Co-Occurrence of Thesaurus

作　　者：蒋庆朝 徐硕[1] 陈孟婕[1] 王立华[1] JIANG Qing-zhao;XU Shuo;CHEN Meng-jie;WANG Li-hua(Institute of Fisheries Engineering of Chinese Academy of Fishery Sciences,Beijing 100141)

机构地区：[1]中国水产科学研究院渔业工程研究所,北京100141

出　　处：《现代计算机》2020年第5期7-11,共5页Modern Computer

基　　金：渔业工程研究所基本科研业务费(No.2019HY-ZC001-2)。

摘　　要：内容推荐方法难以发掘用户潜在兴趣,并且对融合多种主题和关键词的物品推荐程度较低。针对这一问题,提出一种基于主题词频共现的推荐算法,并进行了验证。通过对比主题词频共现的推荐算法与内容推荐、协同过滤、人口统计的推荐结果,发现相较于其他推荐方法,所提出的方法具有更高的准确率和击中率。同时,推荐结果能够体现多种主题融合,并且可以发掘用户感兴趣的主题,弥补推荐结果主题过于集中的缺陷。Nowadays,the content-based recommendation method is difficult to discover the potential interests of users,and the recommendation degree of the articles which combine topics and keywords is low.To solve this problem,this paper proposes a recommendation algorithm based on frequent co-occurrence of topics.And it was validated in National Agricultural Science Data Sharing Center-Fishery Science Data Sharing Center.Firstly,we extract the keywords from users’browsing documents.Then,a list of keyword phrases is obtained through combination.It can get the recommended results by calculating the co-occurrence frequency of key phrases in documents.To build a recommendation system for Fishery Science Data Center and recommend for users.This paper compares the frequency co-occurrence algorithm with content-based recommendation,collaborative filtering and statistical algorithm.The results show that the accuracy and hit rate of the recommendation results based on co-occurrence of topics are higher.The recommendation results reflect the integration of multiple topics and show more documents of keyword phrases.The results prove that this method is more accurate than the traditional algorithm.And it can discover more topics that users may be interested in and makes up for the defect that the topic of results is too centralized.

关 键 词：词频共现 推荐 主题融合 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]