摘要
基于标签的推荐系统通过研究用户打标签行为,为用户进行个性化推荐,因此用户所打标签的质量影响推荐效果,但目前大部分的研究并未考虑到标签的质量问题。针对标签单词拼写错误问题,论文提出一种基于最近邻的标签修正推荐算法(TCNNB)。该算法首先由Spark的RDD离线计算对数据集进行处理,得到所有标签单词的字母频次之差,然后使用欧式距离找出最接近的单词,即为修正后的标签单词,最后为用户进行个性化推荐。实验结果表明,引入TCNNB算法对标签单词进行修正,使推荐的精准度(召回率和准确率)得到了明显提高,较好地改进了基于标签的推荐系统的推荐效果。
The tag recommender system can personalize recommendations for users by studying the tagging behavior of the user,so the quality of tags tagged by users affects the recommendation effect.However,most of the current studies does not consider the quality of labels.For words spelling errors in tags,a Tag Correction Recommendation Algorithm Based on Nearest Neighbor(TCNNB)is proposed in this paper.Firstly,the data set is processed by RDD offline calculation of Spark,then the frequency difference of letters of all words of tags is gotten.Secondly,the closest word is found,which is the corrected word of a tag according to the distance.Finally,based on the tag processed by TCNNB algorithm,personalized recommendation is made for users.The experimental results show that the introduction of TCNNB algorithm to correct the word of tags makes recommendation accuracy(Recall rate and Precision)get more obvious enhancement,and improves the recommendation effect which is based on the tag recommendation system.
作者
余利国
丁卫平
景炜
YU Liguo;DING Weiping;JING Wei(School of Information Science and Technology,Nantong University,Nantong 226019)
出处
《计算机与数字工程》
2020年第4期735-740,共6页
Computer & Digital Engineering
基金
国家自然科学基金项目(编号:61976120)
江苏省自然科学基金项目(编号:BK20191445)
江苏省六大人才高峰项目(编号:XYDXXJS-048)
江苏高校“青蓝工程”(编号:苏教师〔2019〕3号)资助
关键词
推荐系统
标签
最近邻算法
拼写错误修正
recommendation system
tags
nearest neighbor algorithm
spelling correction
