Topic modeling in density functional theory on citations of condensed matter electronic structure packages

Area of Science:

• Computational Chemistry
• Bibliometrics
• Data Science

Background:

• The rapid growth of scientific publications poses challenges for researchers staying current.
• Accurate paper classification is crucial for personalized research discovery.
• Density Functional Theory (DFT) is a widely used methodology with a prolific publication record.

Purpose of the Study:

• To develop an unsupervised method for classifying scientific publications.
• To demonstrate the method's utility using a dataset of DFT-related articles.
• To analyze publication trends and interconnections within a research field.

Main Methods:

• Unsupervised topic modeling based on abstract text analysis.
• Word analysis and clustering to identify publication topics.
• Attributing papers to topics via word similarity.

Main Results:

• Successfully classified DFT-related publications into distinct topics.
• Identified correlations between topics and publication metadata (publishers, journals, country, year).
• The method provides a general framework for analyzing publication trends.

Conclusions:

• The proposed unsupervised topic modeling approach effectively classifies scientific literature.
• This method facilitates better understanding of research landscapes and trends.
• The approach is adaptable for analyzing other scientific disciplines beyond DFT.