Review on Performance Of SDAE For Historical Usage Data Using Deep Learning
Meenakshi, Satpal · 2020
Today, there are huge number of online web-services and size of users have increased in large scaled over the past years. In long-term web services are important to increase the web API economy, how to recommend the long-tail web services efficiently is a primary issue. Moreover, to focus on this problem, the traditional web based recommendation services performs poorly on long-tail side. To overcome the problem of severe sparsity of historical usage data and unsatisfactory quality of description content, we are focusing on Convolutional Neural Network approach to boost the performance of the network and reduce time consumption. A client is an individual who uses the recommendation engine giving his viewpoint about different products and get suggestion about the novel products through the recommender. Generally in recommendation system, user gives input to the recommender. The inputs may be both user defined and built-in inputs. To overcome the sparsity problem and find out 75% accuracy, we use DLSTR methodology by using SDAE.