Research on Simulation of Interactive Behavior of Cognitive Agents Based on Large Language Models (LLMs) and NetLogo

This dataset was constructed to support the research paper titled "Simulation of Cognitive Agent Interaction Behaviors Based on Large Language Models and NetLogo". It is primarily used to replicate the generation of social media cognitive agents and the simulation of their interaction behaviors within a cognitive warfare context, specifically centered on the "Al-Ahli Hospital Explosion Incident".The core source of the dataset is publicly available social media textual data related to the incident, collected from three platforms: Weibo, Twitter, and Douyin. Data collection was performed using Python's Selenium automation testing framework to simulate user browsing behavior, combined with the Requests library for asynchronous data crawling. The collection period spanned from October 17, 2023 (the day of the incident) to March 26, 2024. The raw data underwent a cleaning process, which included the removal of duplicates, content from advertisement bots, and posts unrelated to the topic. This process yielded 39,754 records of cognitive contention content, 29,381 records of basic user information for users involved in the cognitive contention, and 377,353 records of historical information from these users.