Comparative Opinion Quintuple Extraction from Product Reviews

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 3955–3965 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 3955 Comparative Opinion Quintuple Extraction from Product Reviews Ziheng Liu∗ Rui Xia∗† Jianfei Yu School of Computer Science and Engineering, Nanjing University of Science and Technology, China {zhliu, rxia, jfyu}@njust.edu.cn Abstract As an important task in opinion mining, com- parative opinion mining aims to identify com- parative sentences from product reviews, ex- tract the comparative elements, and obtain the corresponding comparative opinion tuples. However, most previous studies simply re- garded comparative tuple extraction as com- parative element extraction, which ignored the fact that many comparative sentences may contain multiple comparisons. The compar- ative opinion tuples defined in these studies also failed to explicitly provide comparative preferences. To address these issues, in this work we first introduce a new Comparative Opinion Quintuple Extraction (COQE) task, to identify comparative sentences from prod- uct reviews and extract all comparative opin- ion quintuples (Subject, Object, Comparative Aspect, Comparative Opinion, Comparative Preference). Secondly, based on the existing comparative opinion mining corpora, we make supplementary annotations and construct three datasets for the COQE task. Finally, we bench- mark the COQE task by proposing a new multi-stage neural network approach which significantly outperforms the baseline systems extended from previous comparative opinion mining methods. The datasets and source code are publicly released at https://github. com/NUSTM/COQE. 1 Introduction Fine-grained opinion mining from product re- views has received considerable attention in the last decade. As around 10% of product reviews contain at least one comparison (Kessler and Kuhn, 2013), it is therefore crucial to extract and analyze these comparative sentences to detect pub- lic opinions towards the compared entities and as- pects. ∗Equal contribution. † Corresponding author. As the pioneering work for this direction, Jin- dal and Liu (2006b) proposed the Comparative Sentence Mining (CSM) task which first identifies comparative sentences from reviews, and extracts pre-defined comparative quintuples, i.e., (Subject, Object, Comparative Aspect, Relation Word, Com- parison Type), from the identified comparative sen- tences. For example, given a sentence “G6 has a worse zoom than G7, but G6’s battery was more reliable than G7”, one comparative quintuple is (G6, G7, battery, more, Non-Equal Gradable). In their work they assumed that a comparative sen- tence contains only one comparative relation, and treated the comparative quintuple extraction task as a comparative element extraction (CEE) prob- lem. However, in real scenarios, a large number of comparative sentences contain more than one com- parative relation. For example, 17.6% of the com- parative sentences in the camera domain (Kessler and Kuhn, 2014) have at least two comparative relations. In this situation, simply applying CEE cannot extract comparative quintuples effectively. Moreover, in their definition, the relation word sometimes fails to explicitly reflect the compara- tive preference. For example, “more” in the above sentence is ambiguous, since it may refer to “more reliable” or “more expensive”. Some recent studies (Panchenko et al., 2019; Ma et al., 2020) proposed a new task named Com- parative Preference …

Content truncated to 500 words