Abstract

Web services and its tremendous applications in many fields provide a platform to the users to meet their requirements thus giving an opportunity to leverage its benefits in many ways. Mash up design in web has become popular in predicting web service interactions and compositions. In this paper, we implement two deep learning techniques to classify web services: stacked denoising autoencoder and convolutional neural network. By classifying web API, we can predict the future interactions of each service in the service network and help the developers for perfect service selection and also in mash up designing. We trained a stacked denoising autoencoder for learning the latent features of web services and deployed a convolutional neural network for classification of web services for predicting future service interactions. Experimental results show that deep learning achieves good performance and error optimization in web API classification.

Paper Details