Smart waste classification in IoT-enabled smart cities using VGG16 and Cat Swarm Optimized random forest.

Area of Science:

• Computer Science
• Environmental Science
• Artificial Intelligence

Background:

• Sustainable urban development increasingly relies on effective waste management.
• Smart cities leverage IoT technologies to enhance urban services, including waste management.
• Efficient garbage sorting and recycling are critical for environmental sustainability.

Purpose of the Study:

• To develop and evaluate a novel waste categorization model for smart city applications.
• To enhance garbage sorting and recycling mechanisms using smart technologies.
• To compare the performance of the proposed model against conventional machine learning approaches.

Main Methods:

• Utilized transfer learning with the VGG16 model for feature extraction from waste images.
• Implemented a Random Forest classifier optimized by Cat Swarm Optimization (CSO).
• Trained and tested the model on a Kaggle garbage categorization dataset.

Main Results:

• The VGG16-CSO-Random Forest model achieved 85% accuracy and an AUC of 0.85.
• Outperformed traditional models such as Support Vector Machines (SVM), XGBoost, and logistic regression.
• Demonstrated superior performance in precision, recall, and F1-score.

Conclusions:

• The proposed waste categorization model shows significant potential for smart city waste management.
• Transfer learning combined with optimized Random Forest offers an effective solution for automated garbage sorting.
• This approach can contribute to more efficient recycling and sustainable urban environments.