Aspect-based sentiment analysis (ABSA) aims at identifying fine-grained polarity of opinion associated with a given aspect word. Several existing articles demonstrated promising ABSA accuracy using ...positional embedding to show the relationship between an aspect word and its context. In most cases, the positional embedding depends on the distance between the aspect word and the remaining words in the context, known as the position index sequence. However, these techniques usually employ both complex preprocessing approaches with additional trainable positional embedding and complex architectures to obtain the state-of-the-art performance. In this paper, we simplify preprocessing by including polarity lexicon replacement and masking techniques that carry the information of the aspect word’s position and eliminate the positional embedding. We then adopt a novel and concise architecture using two Bidirectional GRU along with an attention layer to classify the aspect based on its context words. Experiment results show that the simplified preprocessing and the concise architecture significantly improve the accuracy of the publicly available ABSA datasets, obtaining 81.37%, 75.39%, 80.88%, and 89.30% in restaurant 14, laptop 14, restaurant 15, and restaurant 16 respectively.
•Proposes an unsupervised method of product feature extraction by using comparative domain corpora.•A novel term similarity measure is introduced to evaluate the association of candidate features and ...domain entities.•Proposes feature-oriented opinion lexicon generation for feature-oriented opinion determination.
Identifying product features from reviews is the fundamental step as well as a bottleneck in feature-level sentiment analysis. This study proposes a method of unsupervised product feature extraction for feature-oriented opinion determination. The domain-specific features are extracted by measuring the similarity distance of domain vectors. A domain vector is derived based on the association values between a feature and comparative domain corpora. A novel term similarity measure (PMI–TFIDF) is introduced to evaluate the association of candidate features and domain entities. The results show that our approach of feature extraction outperforms other state-of-the-art methods, and the only external resources used are comparative domain corpora. Therefore, it is generic and unsupervised. Compared with traditional pointwise mutual information (PMI), PMI–TFIDF showed better distinction ability. We also propose feature-oriented opinion determination based on feature-opinion pair extraction and feature-oriented opinion lexicon generation. The results demonstrate the effectiveness of our proposed method and indicate that feature-oriented opinion lexicons are superior to general opinion lexicons for feature-oriented opinion determination.
Vital to the task of Sentiment Analysis (SA), or automatically mining sentiment expression from text, is a sentiment lexicon. This fundamental lexical resource comprises the smallest ...sentiment-carrying units of text, words, annotated for their sentiment properties, and aids in SA tasks on larger pieces of text. Unfortunately, digital dictionaries do not readily include information on the sentiment properties of their entries, and manually compiling sentiment lexicons is tedious in terms of annotator time and effort. This has resulted in the emergence of a large number of research works concentrated on automated sentiment lexicon generation. The dictionary-based approach involves leveraging digital dictionaries, while the corpus-based approach involves exploiting co-occurrence statistics embedded in text corpora. Although the former approach has been exhaustively investigated, the majority of works focus on terms. The few state-of-the-art models concentrated on the finer-grained term sense level remain to exhibit several prominent limitations, e.g., the proposed semantic relations algorithm retrieves only senses that are at a close proximity to the seed senses in the semantic network, thus prohibiting the retrieval of remote sentiment-carrying senses beyond the reach of the ‘radius’ defined by number of iterations of semantic relations expansion. The proposed model aims to overcome the issues inherent in dictionary-based sense-level sentiment lexicon generation models using: (1) null seed sets, and a morphological approach inspired by the Marking Theory in Linguistics to populate them automatically; (2) a dual-step context-aware gloss expansion algorithm that ‘mines’ human defined gloss information from a digital dictionary, ensuring senses overlooked by the semantic relations expansion algorithm are identified; and (3) a fully-unsupervised sentiment categorization algorithm on the basis of the Network Theory. The results demonstrate that context-aware in-gloss matching successfully retrieves senses beyond the reach of the semantic relations expansion algorithm used by prominent, well-known models. Evaluation of the proposed model to accurately assign senses with polarity demonstrates that it is on par with state-of-the-art models against the same gold standard benchmarks. The model has theoretical implications in future work to effectively exploit the readily-available human-defined gloss information in a digital dictionary, in the task of assigning polarity to term senses. Extrinsic evaluation in a real-world sentiment classification task on multiple publically-available varying-domain datasets demonstrates its practical implication and application in sentiment analysis, as well as in other related fields such as information science, opinion retrieval and computational linguistics.
Aspect-based sentiment analysis (ABSA) includes aspect extraction and sentiment analysis on those extracted aspects. This paper is mainly focused on the explicit aspect extraction from product based ...online customer reviews. Many research studies on explicit aspect extractions have adopted dependency rule-based techniques but limited their focus to only nouns and noun phrases as potential aspects. Moreover, extraction of multiple-aspects and multi-word aspects from online reviews are also not addressed by previous research studies. In this paper, we have proposed a Dependency structure-based rules using ROOT Node (DS-RN) technique using spaCy dependency parser to extract nouns, noun phrases, verbs, verb phrases in addition to single word, multiple aspects and multi word aspect extractions from customer review datasets. In our proposed methodology, 30 new dependency-based rules are formulated and implemented on 5 different product-based datasets. This study is based on the pattern analysis of dependency structures of review sentences to develop dependency-based rules for explicit aspect extraction. The proposed approach also incorporated lexicon-based pruning techniques to remove irrelevant aspects and retain correct aspects. The performance results on 5 different product-based customer review datasets demonstrate that our proposed DS-RN approach outperforms all other state-of-the-art baseline works with averaged value of precision as 87%, recall with 97% and 91% as F1-score.
With the rapid development of web, a huge number of reviews about various kinds of products are springing on the Internet. Many users purchase the products through online thereby reduce the time ...consumption by avoid travelling. An opinion lexicon is a file of opinion words which is used to indicate the list of positive, negative or neutral sentiments in the reviews. However it is difficult to extract features from a large corpus, so identifying the relations between opinion words and opinion targets of a specific domain allows thorough understanding of customers opinion. In existing system, identifying opinion relation between opinion words and opinion targets are performed by using word alignment model. Identifying relation among words is done by graph-based co-ranking algorithm which is used to estimate the confidence of each candidate. To filter false candidates we propose a novel framework for fine-grained opinion mining where involves propagation and refinement process. This is used to filter false candidates for effective opinion to the users by using three-layer opinion relation graph to identify potential relations and to rank all the feature candidates which effectively alleviates the problems of error propagation and infrequent results discovering. For attaining a stable syntactic pattern set, propagation and refinement process is done and the final review text is considered to be an opinion of specific product. The performance of proposed system outperforms the framework of fine-grained opinion mining by reducing the error propagation and false results are removed from the online review.
The complexity of comments from the Senegalese on-line press is mainly due to ambiguity. This has led to customized abbreviations and the high presence of local languages. These factors make the ...current opinion mining tools ineffective. In view of this complexity, our objective is to suggest an opinion lexicon to process these types of data. This lexicon will take into account Wolof, French and urban language words and expressions. The significance of such a study is to provide our local languages with tools for automatic processing of natural languages.
Aspect-based sentiment analysis (ABSA) includes aspect extraction and sentiment analysis on those extracted aspects. This paper is mainly focused on the explicit aspect extraction from product based ...online customer reviews. Many research studies on explicit aspect extractions have adopted dependency rule-based techniques but limited their focus to only nouns and noun phrases as potential aspects. Moreover, extraction of multiple-aspects and multi-word aspects from online reviews are also not addressed by previous research studies. In this paper, we have proposed a Dependency structure-based rules using ROOT Node (DS-RN) technique using spaCy dependency parser to extract nouns, noun phrases, verbs, verb phrases in addition to single word, multiple aspects and multi word aspect extractions from customer review datasets. In our proposed methodology, 30 new dependency-based rules are formulated and implemented on 5 different product-based datasets. This study is based on the pattern analysis of dependency structures of review sentences to develop dependency-based rules for explicit aspect extraction. The proposed approach also incorporated lexicon-based pruning techniques to remove irrelevant aspects and retain correct aspects. The performance results on 5 different product-based customer review datasets demonstrate that our proposed DS-RN approach outperforms all other state-of-the-art baseline works with averaged value of precision as 87%, recall with 97% and 91% as F1-score.
The services of Indian Railway are availed by many people in the country. It is an important mode of transportation. Most of the users of Indian Railway express their views about it on different ...social media sites like Twitter, Facebook etc. It leads to generation of large amount of data and sentimental analysis of that data can be very helpful in understanding public opinions towards Indian Railway and in decision making. In this paper, the lexicon based sentimental analysis technique has been applied to the twitter data collected corresponding to three train accidents namely Puri-Haridwar-Kalinga Utkal Express, Delhi-bound Kaifiyat Express and Mumbai-Nagpur Duranto Express which took place on 19/08/2017, 23/08/2017 and 29/08/2017 respectively. Further, tweets are classified into different categories and analyzed in terms of percentage frequency. The results present the pattern how the sentiments of the public fluctuate with time as when derailment happens the negative tweets has high frequency of occurrence but with passage of time frequency of occurrence of neutral tweets become high.
Opinion analysis has become important since there're huge amount of opinions and comments in user-generated content for social media such as blogs and forums. While machine learning methods classify ...opinions by statistical properties such as term frequency, lexicon-based methods could reflect the intricate semantics in determining the opinion orientation and strength. In this paper, we propose an unsupervised opinion phrase extraction and rating mechanism with opinion lexicons. First, opinion phrases are extracted from each blog post by matching terms in the opinion lexicon. Then, scores are assigned to each opinion phrase for different scales of opinion strength. Finally, the total score for the post is estimated by averaging all of its opinion phrases. From the experimental results on digital camera posts, the proposed method achieved high accuracy with sigmoid scoring. This shows effective extraction of opinion phrases and rating of opinion scales. Further investigation is needed to test the effectiveness in different document domains.