A comprehensive understanding of the multifaceted phenomenon of terrorism requires a multidisciplinary approach integrating psychology, data science, and systems modeling. This study proposes a novel method for detecting psychological indicators of terrorism using a keyword categorization technique based on importance and semantic proximity, enabling user-defined subgroups. Another objective was to process large datasets with minimal resources, enhance keyword efficiency, and simplify literature analysis. To this end, a simulation-based application was designed to examine the effects of psychological dimensions on terrorism by combining large-scale literature analysis with system dynamics modeling.


An extensive corpus was compiled by collecting abstracts from 7,244 unique articles published between 1976 and 2023 (5,295 from Web of Science and 1,949 from Scopus). After preprocessing, the dataset consisted of 1,126,722 word tokens. In order to identify psychological indicators relevant to terrorism, 9,456 keywords were extracted using Shannon's entropy difference. A novel clustering score formula was then implemented, integrating entropy difference and cosine similarity scores, yielding four heterogeneous submatrices -experts, novices, public figures, and the general public- each containing 20-24 million keyword pairs and capturing symmetrical associations among psychological constructs.


Network analysis was conducted to identify main and sub-topics associated with terrorism. The results revealed clusters of psychological concepts strongly associated with terrorism, including social polarization, absolutism, threat orientation, hate, violence, and discrimination. These dimensions were treated as variables within a system dynamics framework. Based on these findings, an interactive simulation application was developed to model the influence of psychological factors and their interactions on terrorism.


The application contributes methodologically by linking text-mined constructs with dynamic simulation, offering a replicable and resource-efficient tool for analyzing complex phenomena. This study highlights the practical relevance of systematically exploring psychological factors shaping terrorism and provides a platform for scenario-based training and policy analysis.