Lifelong-RL: Lifelong Relaxation Labeling for Separating Entities and Aspects in Opinion Targets

Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 225–235, Austin, Texas, November 1-5, 2016. c⃝2016 Association for Computational Linguistics Lifelong-RL: Lifelong Relaxation Labeling for Separating Entities and Aspects in Opinion Targets Lei Shu1, Bing Liu1, Hu Xu1, Annice Kim2 1Department of Computer Science, University of Illinois at Chicago, USA 2Center for Health Policy Science and Tobacco Research, RTI International, USA 1{lshu3, liub, hxu48}@uic.edu, 2akim@rti.org Abstract It is well-known that opinions have targets. Extracting such targets is an important prob- lem of opinion mining because without know- ing the target of an opinion, the opinion is of limited use. So far many algorithms have been proposed to extract opinion targets. However, an opinion target can be an entity or an as- pect (part or attribute) of an entity. An opinion about an entity is an opinion about the entity as a whole, while an opinion about an aspect is just an opinion about that specific attribute or aspect of an entity. Thus, opinion targets should be separated into entities and aspects before use because they represent very dif- ferent things about opinions. This paper pro- poses a novel algorithm, called Lifelong-RL, to solve the problem based on lifelong ma- chine learning and relaxation labeling. Ex- tensive experiments show that the proposed algorithm Lifelong-RL outperforms baseline methods markedly. 1 Introduction A core problem of opinion mining or sentiment anal- ysis is to identify each opinion/sentiment target and to classify the opinion/sentiment polarity on the tar- get (Liu, 2012). For example, in a review sen- tence for a car, one wrote “Although the engine is slightly weak, this car is great.” The person is posi- tive (opinion polarity) about the car (opinion target) as a whole, but slightly negative (opinion polarity) about the car’s engine (opinion target). Past research has proposed many techniques to extract opinion targets (we will just call them targets hereafter for simplicity) and also to classify senti- ment polarities on the targets. However, a target can be an entity or an aspect (part or attribute) of an en- tity. “Engine” in the above sentence is just one as- pect of the car, while “this car” refers to the whole car. Note that in (Liu, 2012), an entity is called a general aspect. For effective opinion mining, we need to classify whether a target is an entity or an as- pect because they refer to very different things. One can be positive about the whole entity (car) but neg- ative about some aspects of it (e.g., engine) and vice versa. This paper aims to perform the target classi- fication task, which, to our knowledge, has not been attempted before. Although in supervised extraction one can annotate entities and aspects with separate labels in the training data to build a model to extract them separately, in this paper our goal is to help un- supervised target extraction methods to classify tar- gets. Unsupervised target extraction methods are of- ten preferred because they save the time-consuming data …

Content truncated to 500 words