A Cross-Cultural Leadership Behavior Prediction Model for Advancing Organizational and Global Management Practices

• Mohanty, Ashish ^[2] ; Ahmad Wani, Tanveer ^[1] ; S, Varalakshmi ^[3] ; Homavazir, Zuleika ^[4] ; Singh, Ankita ^[5] ; M, Lavanya ^[6]
  1. [1] Noida International University

     Noida International University

     IN.36.141.7279602, India

     
  2. [2] Department of Management, Institute of Business and Computer Studies, Siksha 'O' Anusandhan (Deemed to be University), Bhubaneswar, Odisha, India
  3. [3] Department of Management, Jain (Deemed to be University), Bangalore, Karnataka, India
  4. [4] ISME, ATLAS SkillTech University, Mumbai, India
  5. [5] Department of Commerce, ARKA JAIN University, Jamshedpur, Jharkhand, India
  6. [6] Master Of Business Administration, Sathyabama Institute of Science and Technology, Chennai, India

  Mostrar afiliaciones +
• Localización: Management: (Montevideo), ISSN-e 3046-4048, Vol. 3, Nº. 0, 2025 (Ejemplar dedicado a: Management (Montevideo))
• Idioma: inglés
• Enlaces
  • Texto completo (pdf)
• Resumen

  • In today’s globally interconnected environment, recognizing culturally preferred leadership behaviors is dynamic for effective organizational and global management. This research presents a cross-cultural leadership behavior prediction model that employs advanced machine learning (ML) techniques to analyze and forecast leadership preferences across diverse cultural contexts. The research gathers data from Kaggle, comprising 6,945 responses across five countries, to analyze behavioral, demographic, and cultural aspects, and employs preprocessing techniques for improved model reliability. The model employs algorithms, such as the Scalable Golden Jackal Optimizer-driven Stacked Random Forest (SGO-SRF) to predict and uncover patterns in leadership behavior preferences. Cultural indicators and demographic features are analyzed using Recursive Feature Elimination (RFE) to identify their impact on these various leadership dimensions. The model was primarily applied using a baseline Random Forest (RF), then established through a Stacked RF approach, and finally optimized using the proposed hybrid SGO-SRF, which attained the highest performance across all evaluation metrics. The hybrid method was implemented in Python and it demonstrated superior performance, achieving higher accuracy (92.11%), F1-scores (88.09%), precision (81.98%), and Recall (85.19%). The research reveals that cultural values significantly influence leadership preferences, with long-term orientation affecting uncertainty tolerance, restraint affecting integration, and entrepreneurial values influencing structure, production emphasis, and predictive performance.