Leveraging Machine Learning for Predictive Analysis in Incomplete Data Scenarios

The workplace conditions, such as optimal CO
2
level and humidity are essential factors to ensure the safety of employees. The work safety concerns established practices to protect employees' health and ensure wellbeing. Studies show that various human and environmental factors impact infections spreading across the office space. By monitoring these factors a less infectious workplace can be created. To reach this goal, an IoT enabled solution DROVIDS with continuous environment monitoring capabilities is proposed. The solution works with near real-time data and integrates predictive analysis, allowing to gain a comprehensive understanding of the environment and detect risks precisely. The solution enables near real-time response and provides a straightforward and immediate operational setup. The DROVIDS analytical service offers adjustability, it can be used without the missing data prediction component if there is no intention of calculating risk and receiving possible mitigations for future planned events. This paper presents DROVIDS architecture, design and operational principles.