| Abstract |
This study introduces an innovative framework using Long Short-Term Memory (LSTM) neural networks to analyze the impact of digital leadership on employee productivity. Motivated by the challenges of interpreting complex organizational data, the research applies a 24-month longitudinal study of 150 organizations and 2.8 million data points, achieving 92.3% accuracy in identifying leadership effectiveness patterns. The LSTM model, enhanced with bidirectional layers and advanced regularization techniques, demonstrates a 23.7% improvement in processing efficiency. Findings reveal three temporal impact clusters: immediate (0-48 hours), medium-term (3-7 days), and long-term (>2 weeks), with strategic decision-making events having the highest impact weight (0.412). The framework enhances decision-making efficiency by 27.8% and team coordination by 34.2%, supported by an hourly prediction precision of 0.934 and daily accuracy of 0.912 across organizational scales. While this research also highlights potential applications in smart city governance, such as optimizing resource allocation and improving crisis management, its primary contribution lies in advancing AI-driven organizational analytics. By integrating AI methodologies, the framework provides actionable insights into digital leadership dynamics, offering scalable, real-time solutions to enhance productivity and decision-making in the digital age. © 2025 The Authors. |
