Integrating word status for joint detection of sentiment and aspect in reviews

Abstract

A crucial task in sentiment analysis is aspect detection: the step of selecting the aspects on which opinions are expressed. This step anticipates the step of determining whether the opinions on aspects are positive or negative. This article proposes a novel probabilistic generative topic model for aspect-based sentiment analysis which is able to discover the latent structure of a large collection of review documents. The proposed joint sentiment-aspect detection model (SAM) is a generative topic model that incorporates the structure of review sentences for detecting aspects and sentiments simultaneously. The intuitions behind the SAM are that from generating documents by latent single- and multi-word topics, modelling the word distribution for each topic and learning of the prior distribution over topics in sentences of documents. SAM introduces word status so that the model can decide when to sample from a bigram distribution or a unigram distribution and integrates all these components into one combined model for aspect-based sentiment analysis. We evaluate SAM both qualitatively and quantitatively to show that the model is indeed able to perform the task effectively and improves significantly over standard joint sentiment-aspect models. The proposed model can easily be transformed between domains or languages and can detect the polarity of text data at various levels. However, for the quantitative analysis, we mainly focus on presenting the results for the document-level sentiment classification.

A Combination of Frequent Pattern Mining and Graph Traversal Approaches for Aspect Elicitation in Customer Reviews

“…We evaluate our approach in extraction of multi-word aspects separately. Figure 16 depicts the comparison of our approach to four methods of LRT-based [3], SAM [4], ELDA [8] and MMI-based [14] in extracting three forms of multi- word aspects since the aim of these approaches are extraction multi-word aspects in Persian language from different perspectives. As could be seen, the results of our method outstrip the results of other four approaches in detecting the type C multi-word aspects.…”

“…The author applied SAM on English and Persian languages on three domains of film, cell phone, DVD players' and restaurant reviews. This research listed some obstacles in detecting aspects from Persian sentences as illustrated in table 1 [4].…”

“…Linguistic related challenges in Persian sentences for aspect extraction process [3,4,14] Shams and Baraani (2017) worked on aspect extraction by proposing ELDA that works based on topic modeling algorithm of LDA and utilization of a prior knowledge of those similar aspects that occur together on the related topics in an iterative manner. They evaluated their algorithm on English and Persian languages [8].…”

Sentiment analysis of tweets through Altmetrics: A machine learning approach

“…It has become a notable vehicle for the fast creation, transmission and discourse of news stories [18][19][20]. Entirely due to the help from Twitter, Web 2.0 has evolved into a much bigger platform for marketing campaigns [21,22], research-work discussion and sentiment analysis [23][24][25][26]. Twitter is used by many research communities as a channel to increase visibility and reach broader audiences that are interested in discussing scientific literature [27,28], and the most Altmetrics Social Media Metrics research has either focused on or included Twitter [4].…”

SBTM: A joint sentiment and behaviour topic model for online course discussion forums

“…In this case, the distribution of sentiments for learners’ focused topics should also be discerned. Previous studies [4 –7] have shown that compared with the single topic model latent Dirichlet allocation (LDA), the unified language models of topic and sentiment present more advantages in guiding the generative process of documents such as blogs and reviews. Furthermore, in course forums, textual content is closely associated with their interactive discussion behaviours (e.g.…”