Forecasting medical waste in Istanbul using a novel nonlinear grey Bernoulli model optimized by firefly algorithm

Area of Science:

• Environmental Science
• Public Health
• Data Science

Background:

• Global waste management is critical, exacerbated by COVID-19 related medical waste concerns.
• Istanbul faces significant challenges in managing escalating volumes of medical waste.
• Accurate medical waste volume estimation is essential for effective resource planning.

Purpose of the Study:

• To estimate medical waste volume in Istanbul using a novel hybrid model.
• To introduce and validate the Firefly Algorithm-optimized Nonlinear Grey Bernoulli Model (FA-NGBM(1,1)).
• To compare the proposed model's performance against existing prediction algorithms.

Main Methods:

• Development of the FA-NGBM(1,1) hybrid model incorporating a rolling mechanism and parameter optimization.
• Comparative analysis with classical GM(1,1), FA-GM(1,1), FA-FGM(1,1), and linear regression models.
• Validation using testing and validation datasets to assess prediction accuracy.

Main Results:

• The FA-NGBM(1,1) model achieved a Mean Absolute Percentage Error (MAPE) of 3.47% for testing data and 2.57% for validation data.
• The proposed hybrid model demonstrated superior accuracy compared to all other evaluated prediction algorithms.
• Forecasts indicate a substantial increase in Istanbul's medical waste over the next three years.

Conclusions:

• The FA-NGBM(1,1) model offers a highly accurate and reliable method for medical waste volume estimation.
• The study highlights the urgent need for enhanced waste management policies in Istanbul.
• Proactive measures are recommended for decision-makers to address the projected rise in medical waste.