Hybrid Particle Swarm Optimization Multi Layer Perceptron for Web-Services Classification

Abstract

The Web services are applications that perform specific tasks and are accessible via the network through a communication protocol. QoS plays an important role in finding out the performance of web services. Multi-layer perceptron neural network (MLP) is the most popular and widely used nonlinear network for solving many practical problems in applied science, biology, and engineering. In this paper, hybrid Particle Swarm Optimization (PSO) with MLP is performed for medical web-service classification. MLP performance is based on initial weights setting. Conventional training algorithms like Back propagation (BPP) and Levenberg-Marquardt (LM) have slow convergence and local minima problems. Results show that the proposed method performs better accuracy, average precision, average recall and RMSE.