SHINE: Signed Heterogeneous Information Network Embe dding
for Sentiment Link Prediction
Hongwei Wang
∗
Shanghai Jiao Tong University
Shanghai, China
wanghongwei55@gmail.com
Fuzheng Zhang
Microsoft Research Asia
Beijing, China
fuzzhang@microsoft.com
Min Hou
University of Science and Technology
of China, Hefei, Anhui, China
hmhoumin@gmail.com
Xing Xie
Microsoft Research Asia
Beijing, China
xingx@microsoft.com
Minyi Guo
†
Shanghai Jiao Tong University
Shanghai, China
guo-my@cs.sjtu.edu.cn
Qi Liu
University of Science and Technology
of China, Hefei, Anhui, China
qiliuql@ustc.edu.cn
ABSTRACT
In online social networks people often express attitudes towards
others, which forms massive sentiment links among users. Predict-
ing the sign of sentiment links is a fundamental task in many areas
such as personal advertising and public opinion analysis. Previous
works mainly focus on textual sentiment classication, however,
text information can only disclose the “tip of the iceberg” about
users’ true opinions, of which the most are unobserved but implied
by other sources of information such as social relation and users’
prole. To address this problem, in this paper we investigate how
to predict possibly existing sentiment links in the presence of het-
erogeneous information. First, due to the lack of explicit sentiment
links in mainstream social networks, we establish a labeled het-
erogeneous sentiment dataset which consists of users’ sentiment
relation, social relation and prole knowledge by entity-level sen-
timent extraction method. Then we propose a novel and exible
end-to-end Signed Heterogeneous Information Network Embedding
(SHINE) framework to extract users’ latent representations from
heterogeneous networks and predict the sign of unobserved sen-
timent links. SHINE utilizes multiple deep autoencoders to map
each user into a low-dimension feature space while preserving the
network structure. We demonstrate the superiority of SHINE over
state-of-the-art baselines on link prediction and node recommen-
dation in two real-world datasets. The experimental results also
prove the ecacy of SHINE in cold start scenario.
ACM Reference Format:
Hongwei Wang, Fuzheng Zhang, Min Hou, Xing Xie, Minyi Guo, and Qi
Liu. 2018. SHINE: Signed Heterogeneous Information Network Embedding
for Sentiment Link Prediction. In Proceedings of the 11th ACM International
Conference on Web Search and Data Mining (WSDM’18). ACM, New York,
NY, USA, 9 pages. https://doi.org/10.1145/3159652.3159666
∗
This work is done while H. Wang and M. Hou are visiting Microsoft Research Asia.
†
M. Guo is the corresponding author.
Permission to make digital or hard copies of all or part of this work for personal or
classroom use is granted without fee provided that copies are not made or distributed
for prot or commercial advantage and that copies bear this notice and the full citation
on the rst page. Copyrights for components of this work owned by others than ACM
must be honored. Abstracting with credit is permitted. To copy otherwise, or republish,
to post on servers or to redistribute to lists, requires prior specic permission and/or a
fee. Request permissions from permissions@acm.org.
WSDM’18, February 5–9, 2018, Marina Del Rey, CA, USA
© 2018 Association for Computing Machinery.
ACM ISBN 978-1-4503-5581-0/18/02. . .$15.00
https://doi.org/10.1145/3159652.3159666
1 INTRODUCTION
The past decade has witnessed the proliferation of online social
networks such as Facebook, Twitter and Weibo. In these social
network sites, people often share feelings and express attitudes
towards others, e.g., friends, movie stars or politicians, which forms
sentiment links among these users. Dierent from explicit social
links indicating friend or follow relationship, sentiment links are
implied by the semantic content posted by users, and involve dif-
ferent types: positive sentiment links express like, trust or support
attitudes, while negative sentiment links signify dislike or disap-
proval of others. For example, a tweet saying “Vote Trump!” shows
a positive sentiment link from the poster to Donald Trump, and
“Trump is mad...” indicates the opposite case.
For a given sentiment link, we dene its sign to be positive
or negative depending on whether its related content expresses
a positive or negative attitude from the generator of the link to
the recipient [
14
], and all such sentiment links form a new net-
work topology called sentiment network. Previous work [
6
,
11
,
15
]
mainly focuses on sentiment classication based on the concrete
content posted by users. However, they cannot detect the existence
of sentiment links without any prior content information, which
greatly limits the number of possible sentiment links that could be
found. For example, if a user does not post any word concerning
Trump, it is impossible for traditional sentiment classiers to ex-
tract the user’s attitude towards him because “one cannot make
bricks without straw”. Therefore, a fundamental question is, can
we predict the sign of a given sentiment link without observing its
related content? The solution to this problem will benet a great
many online services such as personalized advertising, new friends
recommendation, public opinion analysis, opinion polls, etc.
Despite the great importance, there is little prior work concern-
ing predicting the sign of sentiment links among users in social
networks. The challenges are two-fold. On the one hand, lack of
explicit sentiment labels makes it dicult to determine the polarity
of existing and potential sentiment links. On the other hand, the
complexity of sentiment generation and the sparsity of sentiment
links make it hard for algorithms to achieve desirable performance.
Recently, several studies [
12
,
14
,
31
,
35
] propose methods to solve
the problem of predicting signed links. However, they rely heavily
on manually designed features and cannot work well in real-world
scenarios. Another promising approach called network embedding
WSDM’18, February 5-9, 2018, Marina Del Rey, CA, USA