TJUdeM: A Combination Classifier for Aspect Category Detection
and Sentiment Polarity Classification
Zhifei Zhang and Jian-Yun Nie
Dept. of Comp. Sci. and Oper. Res.
University of Montreal
Quebec H3C 3J7, Canada
{zhanzhif, nie}@iro.umontreal.ca
Hongling Wang
Dept. of Comp. Sci. and Tech.
Soochow University
Suzhou 215006, China
hlwang@suda.edu.cn
Abstract
This paper describes the system we submitted
to In-domain ABSA subtask of SemEval 2015
shared task on aspect-based sentiment analy-
sis that includes aspect category detection and
sentiment polarity classification. For the as-
pect category detection, we combined an SVM
classifier with implicit aspect indicators. For
the sentiment polarity classification, we com-
bined an SVM classifier with a lexicon-based
polarity classifier. Our system outperforms the
baselines on both the laptop and restaurant do-
mains and ranks above average on the laptop
domain.
1 Introduction
Sentiment analysis aims at identifying people’s
opinions, sentiments, attitudes, and emotions to-
wards entities and their attributes (Liu, 2012), which
has a wide range of applications on user-generated
content, e.g., reviews, blogs, and tweets.
Most previous work in sentiment analysis mainly
attempted to identify the overall polarity of a giv-
en text or text span (Pang and Lee, 2008; Wilson et
al., 2009; Zhang et al., 2009). The document-level
or sentence-level sentiment classification is often in-
sufficient for applications. Each document may talk
about different entities, or express different opinions
about different aspects of the entity even if the doc-
ument concerns a single entity. Therefore, we need
to discover the aspects of entities and determine the
sentiment polarity on each entity aspect. This task
is called aspect-based sentiment analysis or feature-
based opinion mining (Hu and Liu, 2004).
The aspect-based sentiment analysis (ABSA) task
(Task 12) (Pontiki et al., 2015) in SemEval 2015 is
a continuation of SemEval 2014 Task 4 (Pontiki et
al., 2014). The ABSA task consists of two subtasks:
In-domain ABSA and Out-domain ABSA. We par-
ticipated in the former subtask that aims to identify
the aspect category (i.e., an entity and attribute pair)
and the sentiment polarity given a review text about
a laptop or a restaurant.
Each entity and attribute pair is an aspect catego-
ry chosen from the predefined inventories of entity
types and attribute labels per domain. For the aspect
category detection, an SVM classifier with the bag-
of-words features can be used, and this approach is
used as our baseline method. However, if a token
implying an aspect, e.g., “overpriced”, is not tak-
en as a feature, the SVM classifier cannot correct-
ly identify its corresponding category. Therefore,
we enhance the results from the SVM classifier by
using implicit aspect indicators (Cruz-Garcia et al.,
2014). For the sentiment polarity classification, an
SVM classifier with the bag-of-words features plus
the category feature is trained and this is used as our
baseline. However, again, if a sentiment word does
not appear in the training data, the SVM classifier
cannot predict its polarity. Therefore, we combined
the SVM classifier and a lexicon-based polarity clas-
sifier (Taboada et al., 2011).
The remainder of this paper is organized as fol-
lows. In Section 2, we describe our approach to
the aspect category detection. In Section 3, our ap-
proach to the sentiment polarity classification is pre-
sented. Experimental results are shown in Section 4.
Section 5 provides the conclusion.