Topic Analysis of the Research Domain in Knowledge Organization: A Latent Dirichlet Allocation Approach

Joo, Soohyung; Choi, Inkyung; Choi, Namjoo

doi:10.5771/0943-7444-2018-2-170

Cited by 16 publications

(22 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Its performance will be tested by processing real-time information of the Twitter platform via the Twitter API, and more advanced text preprocessing techniques will be incorporated to improve the quality of classification of the sentiment. Further enhancements include the portability to other platforms such as mobile devices and the incorporation of Latent Dirichlet Allocation (LDA) technique to automatically identify the number of attributes of corpus [57][58][59][60].…”

Section: Discussionmentioning

confidence: 99%

Feature Weighting Based on Inter-Category and Intra-Category Strength for Twitter Sentiment Analysis

Wang

Youn

2018

Applied Sciences

View full text Add to dashboard Cite

The rapid growth in social networking services has led to the generation of a massivevolume of opinionated information in the form of electronic text. As a result, the research on textsentiment analysis has drawn a great deal of interest. In this paper a novel feature weighting approachis proposed for the sentiment analysis of Twitter data. It properly measures the relative significanceof each feature regarding both intra-category and intra-category distribution. A new statistical modelcalled Category Discriminative Strength is introduced to characterize the discriminability of thefeatures among various categories, and a modified Chi-square (2)-based measure is employed tomeasure the intra-category dependency of the features. Moreover, a fine-grained feature clusteringstrategy is proposed to maximize the accuracy of the analysis. Extensive experiments demonstrate thatthe proposed approach significantly outperforms four state-of-the-art sentiment analysis techniquesin terms of accuracy, precision, recall, and F1 measure with various sizes and patterns of training andtest datasets.

show abstract

Section: Discussionmentioning

confidence: 99%

Feature Weighting Based on Inter-Category and Intra-Category Strength for Twitter Sentiment Analysis

Wang

Youn

2018

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…This paper adopts a topic modeling approach (a useful strategy in KO for assessing a domain (e.g., Joo, Choi, and Choi 2018)) to automate the first part of the facet analysis process. Topic modeling techniques identify high-level categories or classes (or "facets" (Vickery 2008)).…”

Section: Methodsmentioning

confidence: 99%

Topic Modeling and Facet Analysis of an Emerging Domain: Research Data Management and Data Curation

Sandy

Froehlich

Hudson‐Vitale

et al. 2019

nasko

View full text Add to dashboard Cite

Research data management (RDM) is often seen as the overarching field that permits research data to be managed, and is related to the field of data curation (DC), a subset of digital curation. Together, RDM and DC (RDM/DC) allow information professionals to work with clients and each other to make data available in support of the research enterprise. An emerging area of scholarly communication, RDM/DC represents a rich area of study from the perspective of knowledge organization (KO). This paper explores the following research question: What can facet analysis tell us about the emerging field of RDM/DC? First, the MAchine Learning for LanguagE Toolkit (MALLET) implementation of Latent Dirichlet Allocation (LDA) is used for topic modelling of abstracts of the RDM/DC scholarly literature. A preliminary analysis of this empirical data by the research team yields a number of topics and, when possible, their relevant aspects or contexts. Facet analysis principles are next applied to these results, producing four general facets: Practice, Stakeholders, Resources, and Study of RDM/DC; however, complex notions infused throughout the field such as “services” and “metadata” do not appear outright in the analysis. Each facet is then further explored through logical division, and the resulting system is encoded in Protégé and visualized using WebVOWL. We conclude that the major areas of emphasis in this data-intensive field will be fundamentally of interest to those in LIS, in scholarly communication, and perhaps increasingly, in KO and other fields that manage and make available data of all kinds.

show abstract

“…Studies using LDA in subdomains of LIS have also been conducted, including information retrieval (Chen et al 2017a), knowledge organization (Joo et al 2018), and electronic health records (Chen et al 2017b).…”

Section: Topic Modeling and Ldamentioning

confidence: 99%

Knowledge structure transition in library and information science: topic modeling and visualization

Miyata

Ishita

Fang³

et al. 2020

Scientometrics

View full text Add to dashboard Cite

The purpose of this research is to identify topics in library and information science (LIS) using latent Dirichlet allocation (LDA) and to visualize the knowledge structure of the field as consisting of specific topics and its transition from 2000–2002 to 2015–2017. The full text of 1648 research articles from five peer-reviewed representative LIS journals in these two periods was analyzed by using LDA. A total of 30 topics in each period were labeled based on the frequency of terms and the contents of the articles. These topics were plotted on a two-dimensional map using LDAvis and categorized based on their location and characteristics in the plots. Although research areas in some forms were persistent with which discovered in previous studies, they were crucial to the transition of the knowledge structure in LIS and had the following three features: (1) The Internet became the premise of research in LIS in 2015–2017. (2) Theoretical approach or empirical work can be considered as a factor in the transition of the knowledge structure in some categories. (3) The topic diversity of the five core LIS journals decreased from the 2000–2002 to 2015–2017.

show abstract

Topic Analysis of the Research Domain in Knowledge Organization: A Latent Dirichlet Allocation Approach

Cited by 16 publications

References 13 publications

Feature Weighting Based on Inter-Category and Intra-Category Strength for Twitter Sentiment Analysis

Feature Weighting Based on Inter-Category and Intra-Category Strength for Twitter Sentiment Analysis

Topic Modeling and Facet Analysis of an Emerging Domain: Research Data Management and Data Curation

Knowledge structure transition in library and information science: topic modeling and visualization

Contact Info

Product

Resources

About