A Semi-Supervised Learning Approach for Quality-Based Web Service Classification

Abstract

The "Web Service Classification and Recommendation" project aims to beautify the system of choosing and recommending internet offerings the use of device learning techniques. The venture is split into two primary additives: Classification Model: This model classifies internet offerings into satisfactory categories—Bronze, Silver, Gold, and Platinum—based on key Quality of Service attributes like response time, availability, reliability, and throughput. The version is built the usage of numerous gadgets studying algorithms, inclusive of Decision Trees, Support Vector Machines, Logistic Regression, K-Nearest Neighbors, Naive Bayes, Random Forest, Multi-Layer Perceptrons, and XGBoost. Additionally, Explainable AI is hired to offer transparency and interpretability within the category selections, allowing users to apprehend why a web service falls into a particular satisfactory class. Recommendation Model: Using the K-Nearest Neighbors set of rules, this version recommends the top 10 most applicable web services primarily based on consumer-inputted QoS attributes. It computes the similarity among the enter statistics and present internet services to provide tailor-made tips. The system's person-pleasant interface lets in customers to add internet provider facts, view type consequences, and receive customized hints. The challenge provides an efficient and scalable answer for choosing awesome net services, permitting users to make information-driven choices based totally on performance metrics. With its twin approach—type and advice—this machine improves selection-making, complements user revel in, and supports organizations in optimizing service choice.