Proliferation Of User-generated Content Research Articles

Extracting Explicit and Implicit Aspects Using Deep Learning

The proliferation of user-generated content on social networks and websites has heightened the significance of sentiment analysis, also known as opinion mining, as a critical tool for comprehending people’s attitudes toward various topics. Aspect-level sentiment analysis, which considers specific aspects or features of texts, provides a more comprehensive view of sentiment analysis. The aspect-level approach encompasses both explicit and implicit aspects, where explicit aspects are readily mentioned in texts while implicit aspects are implied or inferred from contextual clues. Despite the significance of implicit aspects in the overall review, previous research has predominantly focused on explicit aspect extraction. Limited attention has been given to the extraction of implicit aspects, despite their potential impact on capturing the complete sentiment picture of texts. Therefore, this study aims to find an aspect extraction solution capable of identifying and extracting both explicit and implicit aspects from texts. This study compares various machine and deep learning models on the SemEval-2014 and SemEval-2016 restaurant datasets. The experimental analysis demonstrates that the proposed Aspect-BiLSTM model emerged as the best-performing model, achieving high accuracy in classifying both explicit and implicit aspects, with 92.9% accuracy for the 2014 and 90.7% accuracy for the 2016 datasets. Notably, the proposed solution was able to capture multiple aspects of texts, making it more robust and versatile. This study highlights the efficacy of the Aspect-BiLSTM model for aspect extraction, which will give valuable insights into the advancement of aspect-level sentiment analysis. Doi: 10.28991/ESJ-2024-08-01-05 Full Text: PDF

Semantics perception and refinement network for aspect-based sentiment analysis

With the proliferation of user-generated content on the web, sentiment analysis of text has become a current research hotspot. Aspect-based sentiment analysis (ABSA) is a type of fine-grained sentiment analysis that aims to identify the sentiment polarities of an aspect given in a context sentence. To date, most ABSA methods are based on recurrent neural networks (RNNs) and attention mechanism. However, RNNs suffer from losing long-distance dependency as the sentence length increases. In contrast, self-attention (SA) concentrates on extracting the dependencies among words regardless of the distance. However, similar to RNNs, SA is not sensitive enough to perceive the local semantics information for the ABSA task. In addition, the attention mechanism employed in ABSA may introduce some noise, which is not conducive to capturing important sentiment expressions. To address these issues, in this paper, we propose a semantics perception and refinement network (SPRN) for sentiment analysis based on aspects. In the SPRN, a novel structure named the dual gated multichannel convolution (DGMCC) is utilized to acquire the aspect-related semantics features of a sentence. Specifically, based on the sentence’s global dependency obtained by the SA-based mechanism as the context information and aspect information over the sentence, a multichannel convolution (MCC) is designed with multiple semantic spaces to extract informative local semantic features. Moreover, a dual refinement gate (DRG) is proposed to strengthen the interaction between the aspect and context at a fine granularity and enhance the noise filter. To verify the effect of the SPRN, we conduct abundant experiments on 5 benchmark datasets with pretrained BERT and GloVe embeddings. The experimental results demonstrate that the SPRN performs better than 14 state-of-the-art ABSA methods in terms of various evaluation indices such as accuracy, marco-F1 and AUROC.

Emotion Recognition by Textual Tweets Classification Using Voting Classifier (LR-SGD)

The proliferation of user-generated content on social media has made opinion mining an arduous job. As a microblogging platform, Twitter is being used to collect views about products, trends, and politics. Sentiment analysis is a technique used to analyze the attitude, emotions and opinions of different people towards anything, and it can be carried out on tweets to analyze public opinion on news, policies, social movements, and personalities. By employing Machine Learning models, opinion mining can be performed without reading tweets manually. Their results could assist governments and businesses in rolling out policies, products, and events. Seven Machine Learning models are implemented for emotion recognition by classifying tweets as happy or unhappy. With an in-depth comparative performance analysis, it was observed that proposed voting classifier(LR-SGD) with TF-IDF produces the most optimal result with 79% accuracy and 81% F1 score. To further validate stability of the proposed approach on two more datasets, one binary and other multi-class dataset and achieved robust results.

Feature distillation network for aspect-based sentiment analysis

A proliferation of user-generated content on the web is fueling research into sentiment analysis for improved extraction of human emotional information. Aspect-based sentiment analysis (ABSA) is currently at the focus, which seeks to predict the sentiment of certain aspects in text. The primary challenge is to recognize the relevant contexts for different aspects. Most prior approaches combining recurrent neural networks and attention mechanisms inevitably introduce extraneous noise and diminish prediction accuracy. Furthermore, the sentiment of some context words varies with different aspects and cannot be inferred from themselves alone, which is another challenge that prevents attention mechanisms from performing properly. In this study, we propose a feature distillation network (FDN) for reducing noise and distilling aspect-relevant sentiment features. A novel double-gate mechanism is designed to implement the interactions between aspects and their corresponding contexts at a fine granularity. We introduce a contextual nonlinear projection layer before the double-gate mechanism to generate aspect-specific word representations, which enables the double-gate mechanism to accurately distinguish between sentiment features of the same context word that corresponds to the different aspects. Experiments show that the FDN achieves state-of-the-art performance and improves accuracy from 1.0 percent to 2.0 percent on all benchmarks for ABSA task.

Location and Trajectory Identification from Microblogs

The rapid development of social networks has resulted in a proliferation of user-generated content (UGC), which can benefit many applications. In this paper, we study the problem of identifying a user’s locations from microblogs, to facilitate effective location-based advertisement and recommendation. Since the location information in a microblog is incomplete, we cannot get an accurate location from a local microblog. As such, we propose a global location identification method, Glitter. Glitter combines multiple microblogs of a user and utilizes them to identify the user’s locations. Glitter not only improves the quality of identifying a user’s location but also supplements the location of a microblog so as to obtain an accurate location of a microblog. To facilitate location identification, Glitter organizes points of interest (POIs) into a tree structure where leaf nodes are POIs and non-leaf nodes are segments of POIs, e.g., countries, cities, and streets. Using the tree structure, Glitter first extracts candidate locations from each microblog of a user which correspond to some tree nodes. Then Glitter aggregates these candidate locations and identifies top-k locations of the user. Using the identified top-k user locations, Glitter refines the candidate locations and computes top-k locations of each microblog. To achieve high recall, we enable fuzzy matching between locations and microblogs. We propose an incremental algorithm to support dynamic updates of microblogs. We also study how to identify users’ trajectories based on the extracted locations. We propose an effective algorithm to extract high-quality trajectories. Experimental results on real-world datasets show that our method achieves high quality and good performance, and scales well.

Exploring the Intraurban Digital Divide Using Online Restaurant Reviews: A Case Study in Franklin County, Ohio

Web 2.0 has yielded a wealth of publicly available data, largely through the popularization of social media and the proliferation of user-generated content (UGC). Thus, the data void in social science research is diminishing as these new data sources are exploited. The discipline of geography, in particular, stands to benefit considerably from geotagged UGC or volunteered geographic information (VGI). In this article, we utilize a form of UGC—restaurant reviews—to evaluate the geography of Web 2.0 in Franklin County, Ohio. We find that restaurant reviews are largely concentrated in relatively few areas of the city, indicating that the impacts of Web 2.0 have not touched down evenly within the central county of the Columbus metropolitan statistical area. We argue that an intraurban digital divide is still apparent and might even be exacerbated by Web 2.0.

Mobile Hurts, but Only for a Minute: The Impact of Mobility and Time on User Generated Content

The way we communicate is evolving. The latest change may be underway now, brought about by two IT-induced phenomena: The proliferation of user-generated content and the emergence of mobile devices as a viable computing platform. The SoLoMo acronym (i.e., Social, Local, Mobile) – as popularized by John Doerr of venture capital firm Kleiner Perkins Caufield & Byers – captures the convergence of social and mobile and represents an important frontier of computing. However, due to the fast pace at which this evolution is taking place, our knowledge lags. We need a far richer understanding of the impact that the changing form factor of our new communication devices has on human behavior. We ask: what is the impact of the posting medium (mobile app or Web site) on the length, timing and tone of user-generated content? We investigate these differences in the context of online hotel reviews contributed on TripAdvisor. Our findings have practical and theoretical significance as they shed early light on an emerging phenomenon in social computing. The growing presence and success of visual communication applications (e.g., Instagram, Pinterest) and short burst communication (e.g., text messaging, Twitter, WhatsApp) suggest that there may an ongoing shift in the way humans communicate. This visual and short burst communication is more public and more immediate than traditional interactions. It is therefore critical that we understand the effect of the changing mode of communication on the content of our communication. Future customer generated content may very well be an aggregation of short, focused comments written by the customer during or shortly following the service encounter; understanding the impact of the medium and the mobile platform on individuals’ own evaluations of their experience will be critical to designing effective rating systems.

How content contributors assess and establish credibility on the web

The proliferation of user-generated content (UGC) is one of the distinguishing characteristics of Web 2.0. Internet users contribute content online through platforms such as blogs, wikis, video sharing sites, and sites that allow user feedback. Yet little is known of the credibility practices of these content contributors. Through phone interviews conducted with 29 online content contributors, this study investigates how content contributors assess credibility when gathering information for their online content creation and mediation activities, as well as the strategies they use to establish the credibility of the content they create. These contributors reported that they engaged in content creation activities such as posting or commenting on blogs or online forums, rating or voting on online content, and uploading photos, music, or video. We found that credibility judgments made when gathering information for online content creation and mediation activities could be grouped into three levels: intuitive, heuristic, and strategy-based. We identified three distinctive ways of establishing credibility that are applied during different phases of content contribution: ensuring credibility during the content creation phase; signaling credibility during the content presentation phase; and reinforcing credibility during the post-production phase. We also discovered that content contributors tend to carry over the strategies they used for assessing credibility during information gathering to their strategies for establishing the credibility of their own content. Theoretical implications for credibility research and practical implications for developing information literacy programs are discussed.

Exploring Consumer Motivations for Creating User-Generated Content

ABSTRACTThe advent of Web 2.0 technologies has enabled the efficient creation and distribution of user-generated content (UGC), resulting in vast changes in the online media landscape. For instance, the proliferation of UGC has made a strong impact on consumers, media suppliers, and marketing professionals while necessitating research in order to understand both the short and long-term implications of this media content. This exploratory study (n = 325) seeks to investigate consumer consumption and creation of UGC and the attitudinal factors that contribute to these actions. The data confirm the established relationship between attitude and behavior and indicate attitude serves as a mediating factor between the use and creation of UGC. With regard to the creation of UGC, the ego-defensive and social functions of attitude were found to have the most explanatory power.