© Springer International Publishing Switzerland 2015
S. Zhang et al. (Eds.): KSEM 2015, LNAI 9403, pp. 483–488, 2015.
DOI: 10.1007/978-3-319-25159-2_43
A Microblog Recommendation Algorithm
Based on Multi-tag Correlation
Huifang Ma
1,2(
)
, Meihuizi Jia
1
, Meng Xie
1
, and Xianghong Lin
1
1
College of Computer Science and Engineering, Northwest Normal University,
Lanzhou Gansu 730070, China
mahuifang@yeah.net
2
Key Laboratory of Intelligent Information Processing of Chinese Academy of Sciences,
Institute of Computing Technology, Chinese Academy of Sciences, Beijing 10085, China
Abstract. In this paper, we present a microblog recommendation algorithm based
on multi-tag correlation. Firstly, a tag retrieval strategy is designed to add tags for
unlabeled users, the initial user-tag matrix is then constructed and user-tag weights
are set. In order to represent user interests accurately, we fully investigate the asso-
ciations between the tags. Both inner and outer correlation between tags are defined
to conquer the problem of sparsity of user-tag matrix. The user interests can then be
decided and microblogs can be recommended to users. Experimental results show
that the algorithm is effective for microblog recommendation.
Keywords: Microblog recommendation · Tag retrieval · User-Tag weight · Tag
correlation
1 Introduction
As a typical representative application of web 2.0, microblog has attracted a great
number of users and rapidly developed in recent years. It is necessary to develop new
algorithms to provide the personalized service for these users. And high quality in-
formation should be accurately pushed based on the user's interest[3].
In this paper, we present a microblog recommendation algorithm based on multi-
tag correlation. First, a novel user tag retrieval strategy is developed to select the
tags for unlabeled users and a user-tag matrix is created to represent the initial weight
of users’ tags. Second, we construct a correlation matrix of multi-tag by investigating
inner and outer correlation between tags. Third, the original user-tag matrix is updated
by correlation matrix of multi-tag to obtain the final weight.
The basic outline of this paper is as follows: Section 2 presents user method. The
experiments and results are demonstrated in Section 3. Lastly, we conclude our paper
in Section 4.
2 Our Approach
2.1 User Tag Retrieval and User-Tag Matrix Construction
If the tagging service is provided by microblog system, the built-in tags can be direct-
ly used. Otherwise, a tag retrieval method is adopted to acquire the personal tags from
the microblog posted by that user.
评论0
最新资源