CoMSeC: A Comparative Analysis of Various Service Classification Techniques

The rapid advancement of Artificial Intelligence (AI) and Machine Learning (ML) has significantly impacted Web Service Classification, a critical task for service discovery, composition, and selection in various applications. Effective service classification improves security, scalability, cost-effectiveness, and other crucial parameters for service selection in specific applications. This paper aims to formulate a novel classification model and conduct a comparative study to analyze the performance of base models with various classical and modern clustering algorithms. We propose a unique approach based on natural language processing (NLP) combining Word2Vec and BERT models to generate high-dimensional embeddings from the service dataset. These embeddings are further processed using Dense Layers for dimensionality reduction and then used as inputs to different clustering algorithms. The results demonstrate that our approach significantly improves the accuracy and efficiency of classification, providing a comprehensive overview of performance across various combinations and highlighting the advantages of our method.