Abstract:For the accurate scientific evaluation and advancement of science, it is crucial to understand the development state of a given disciplinary domain. Existing comprehension methodologies concentrate on quantitatively analyzing broad subject trends without considering the underlying complex status attributes of the words that support and enrich these surface-level trends. Through the perspective of the word role, this study deepens the examination of domain development in a more granular way. We use Word2vec to identify the representative semantic neighbors of a domain-specific feature word from literature. Then, a word status observation model is provided that classifies the role of these representative words as tree structures, including young leaves, dead leaves, roots, and trunk-branches, based on changes in their similarity ranking divergence and information entropy. The static word role provides a new insight into detecting the inter -nal detailed organizational composition and maturing status, while the dynamic role shift helps track historical development changes in the domain. Taking attention in psychology as a case by extracting psychology articles from Microsoft Academic Graph, the empirical study illustrates the results of the observation model and yields intriguing findings. First, the static word role po-sitioning shows that a word's status can be obviously different. Different status indicates areas currently keeping the states of thriving, obsolete, becoming firm supportive pillars (precipitating), and irregularly developing within the domain, respectively. Second, during different development periods, most of the words are maintaining the role of irregularly changing or tending towards precipitation. A certain proportion of words are tending towards prosperity, while only a small set of words are deviating from the prosperity state. The growth of attention-related studies as a whole is well-grown and gradually approaching maturity. Our study further supports researchers' understanding of the domain development status from a more granular perspective of word roles.

Empirical Analysis on a Keyword-Based Semantic System

Comparative Study of Chinese-English Scientific Topic Ranking Correlations: Computer Science Perspective

Evolutionary features of academic articles co-keyword network and keywords co-occurrence network: Based on two-mode affiliation network

Ensemble analysis of topical journal ranking in bioinformatics

How do authors select keywords? A preliminary study of author keyword selection behavior

Tracing the evolution of physics with a keyword co-occurrence network

Keyword-citation-keyword Network: a New Perspective of Discipline Knowledge Structure Analysis.

Research on the Semantic Measurement in Co-word Analysis.

Empirical analysis of collective human behavior for extraordinary events in blogosphere

A Study on Technology Forecasting based on Co-occurrence Network of Keyword in Multidisciplinary Journals

A study on information behavior of scholars for article keywords selection

Understanding the Domain Development Through a Word Status Observation Model.

Learning the evolution of disciplines from scientific literature: A functional clustering approach to normalized keyword count trajectories

Change Summarization of Diachronic Scholarly Paper Collections by Semantic Evolution Analysis

KeyphraseDS: Automatic Generation of Survey by Exploiting Keyphrase Information

Small-world phenomenon of keywords network based on complex network

Three new bibliometric indicators/approaches derived from keyword analysis

Toward a deeper understanding of Visualization through keyword analysis

Semantic-enhanced topic evolution analysis: a combination of the dynamic topic model and word2vec

The evolution of Complexity co-occurring keywords: bibliometric analysis and network approach

Keyword Extraction in Scientific Documents