A Novel Approach Web Services Based Long Tail Web Services Using Deep Neural Network

Abstract

As long-tail services are playing wider role in Web services, most of the developers are composing various web based services into mashups. Developers are increasing an interest for long-tail services, moreover, there are deep studies to address the recommendation problem using long-tail web services. The main Challenges for recommending long-tail services correctly includes unsatisfactory quality of description content and sparsity of historical data. Long term Web API Services are convenient, flexible and efficient way of interacting with customers, deliver businesses and sharing and exchange data over the web. They allow instant and complicated web services accessible to ubiquitous cell phone devices, such as tablets or smart phones. In the base paper, author proposed the DLSTR methodology using deep learning techniques, where author applied the feed forward neural network using Stack auto encoder denoising (SADE) to remove the data sparsity problem and achieved the 75% accuracy, which is unsatisfactory according to the traditional techniques. To overcome, this problem, we proposed the HDLSTTR technique, where we applied deep learning techniques with improved model of CNN (Convolutional Neural network) and GRU (Gated recurrent units) using the same dataset of Stack auto encoder denoising (SADE) to remove the data sparsity problem and achieved expected 95% accuracy with improved results. According to these improved results of long-term web services, the results are satisfactory and display the good results of keywords recommendation.