Constructing domain-dependent sentiment dictionary for sentiment analysis
- PDF / 1,035,127 Bytes
- 14 Pages / 595.276 x 790.866 pts Page_size
- 27 Downloads / 253 Views
(0123456789().,-volV)(0123456789(). ,- volV)
ORIGINAL ARTICLE
Constructing domain-dependent sentiment dictionary for sentiment analysis Murtadha Ahmed1,2
•
Qun Chen1,2 • Zhanhuai Li1,2
Received: 5 September 2019 / Accepted: 26 February 2020 Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Sentiment dictionary is of great value to sentiment analysis, which is used widely in sentiment analysis compositionality. However, the sentiment polarity and intensity of the word may vary from one domain to another. In this paper, we introduce a novel approach to build domain-dependent sentiment dictionary, SentiDomain. We propose a weak supervised neural model that aims at learning a set of sentiment clusters embedding from the sentence global representation of the target domain. The model is trained on unlabeled data with weak supervision by reconstructing the input sentence representation from the resulting representation. Furthermore, we also propose an attention-based LSTM model to address aspect-level sentiment analysis task based on the sentiment score retrieved from the proposed dictionary. The key idea is to weight-down the non-sentiment parts among aspect-related information in a given sentence. Our extensive experiments on both English and Chinese benchmark datasets have shown that compared to the state-of-the-art alternatives, our proposals can effectively improve polarity detection. Keywords Lexicon sentiment dictionary Neural network Sentiment analysis Aspect-level sentiment analysis
1 Introduction Sentiment analysis, also known as opinion mining [20], has received intensive attention in the last few years, due to the popularity of social networks and online applications (e.g., Yelp, Amazon, BeerAdvocate) that allow people to express their opinions toward different entities or topics (e.g., product and service). Of great value to sentiment analysis, sentiment dictionary has been widely used to either directly
& Murtadha Ahmed [email protected] Qun Chen [email protected] Zhanhuai Li [email protected] 1
School of Computer Science, Northwestern Polytechnical University, Xi’an, Shaanxi, People’s Republic of China
2
Key Laboratory of Big Data Storage and Management, Northwestern Polytechnical University, Ministry of Industry and Information Technology, Xi’an, Shaanxi, People’s Republic of China
compute the score of sentence [27] or regularize the output of various learning models (e.g., LSTM) [34]. In the sentiment dictionary, each word is associated with a score in the range [1; 1] that represents its sentimental intensity and orientation. The state-of-the-art techniques for building sentiment dictionary usually begin with manually labeled seed words and then automatically extend the list based on preexisting knowledge derived from lexicon resources (e.g., SentiWordNet). Manually annotated dictionaries introduced two lists (i.e., positive and negative words) [3, 38, 46]. SentiWordNet [2] is built based on the glosses associated with synsets and on vectorial term rep
Data Loading...