Deep2s: Improving Aspect Extraction in Opinion Mining With Deep Semantic Representation

Abstract

Syntactical rule based approaches for aspect extraction, which are free from expensive manual annotation, are promising in practice. These approaches extract aspects mainly through the dependency relations in the surface sentence structures. However, deep and rich semantic information hidden in sentences which can help improve aspect extraction, is difficult for them to capture. In order to address the problem, this paper first proposes to employ Logic Programming to explore the feasibility of deep semantic representation, then proposes Deep2S, a hybrid rule-based method to improve the performance of aspect extraction. Deep2S integrates deep semantic representation such as Abstract Meaning Representation (AMR) with syntactic structure. It can take advantage of the syntactical rules to obtain dependency relations in the surface structure as well as the semantic rules to capture deep semantic information. Our experiments are conducted on eight popular review datasets using two evaluation metrics. Experimental results demonstrate the usefulness of deep semantic representation and the ability of Deep2S to improve the performance of aspect extraction in opinion mining.

Extracting Explicit and Implicit Aspects Using Deep Learning

“…As illustrated in Figure 2, aspect extraction solutions can be classified as either being rule-based, unsupervised, or supervised. [17]. Zainuddin et al (2017) [13] employed a Stanford dependency parser to extract aspects from texts by breaking them down into distinct grammatical components and determining the implicit aspect words or phrases based on the grammatical dependencies extracted.…”

“…Lastly, the rule-based aspect extraction solution proposed by Li et al (2020) [17] adopted a framework based on the Answer Set Programming (ASP) [19] logic programming paradigm. Their solution first modeled the relationships between the words in its input texts using Abstract Meaning Representation (AMR) graphs [20].…”

Artificial intelligence as toolset for analysis of public opinion and social interaction in marketing: identification of micro and nano influencers

“…Effective analysis of public opinion about products and services through AI has become the basis of marketing strategies today. Researchers, using a scientometric approach, thematic and patent search methods, and extended citation, collect a lot of bibliographic records of opinion mining research, in-depth semantic analysis identify intellectual landscapes and recent events, record citation trends at the level of each domain, assignment of thematic categories, keyword matching, a network of joint citation of documents, and essential articles (Li et al, 2020;Jang et al, 2022). The study of algorithmic and linguistic aspects of collecting opinions is of the most significant interest to the Internet community in terms of understanding, quantifying and applying the emotional orientation of texts.…”

“…The study of algorithmic and linguistic aspects of collecting opinions is of the most significant interest to the Internet community in terms of understanding, quantifying and applying the emotional orientation of texts. Based on the analysis of recent thematic trends, the authors concluded that the practical application of intelligent opinion analysis: forecasting market value and analyzing product reviews attracts the most significant attention of the community (Tsapatsoulis and Djouvas, 2019;Li et al, 2020;Jang et al, 2022).…”

“…Modern research confirms that it is not the presence of AI technologies as such in the marketing arsenal of companies, but the effective management of new competitive advantages that they are able to produce that expands the commercial prospects of brands. The market success of the developed innovations and the management of sales processes depend on the effectiveness of AI implementation and continuous improvement of the quality of opinion mining and their semantic recognition and analysis (Jarek and Mazurek, 2019;Li et al, 2020;Jang et al, 2022).…”