Valuation of Svm Kernel Performance in Organic and Non-Organic Waste Classification

In an era of increasing concern for environmental sustainability, waste management remains an important global issue. Efficient waste classification, in particular distinguishing between organic and recyclable materials, is essential for reducing environmental impact. Traditional manual classification methods are often error-prone and inefficient. This research evaluates the performance of SVM models with RBF and Polynomial kernels for waste classification, using SqueezeNet for feature extraction. Datasets from Kaggle were preprocessed and augmented to improve model training. The experimental results show that the SVM model with RBF kernel outperforms the Polynomial kernel in classifying organic and recyclable waste, with an accuracy of 97.9% compared to 97.3% for the Polynomial kernel. This finding underscores the importance of kernel selection and parameter tuning in optimising SVM models for non-linear classification tasks. This research contributes to the development of more efficient and accurate waste classification technologies, promoting better waste management practices. Further research is recommended to explore advanced feature extraction methods and expand the scope of classification to cover a wider range of waste categories.