Towards a Novel Framework for Trust Driven Web URL Recommendation Incorporating Semantic Alignment and Recurrent Neural Network

Recommendation System plays an important role in delivering relevant data to the user. A recommender system is also used to display relevant websites with respect to the user query. As the amount of malicious web pages in World Wide Web is quite enormous, there is a huge probability that the URL might be harmful to the user. This paper proposes a Trust-based URL recommendation technique using Semantic Alignment driven Knowledge aggregation methodology along with Artificial neural network and Glowworm Swarm Optimization. The data used for training the Recurrent Neural Network is the URL Trees formulated from the dataset combined with data after classification from fact-checkers, which is later used to check the Threat level of the URL from the initial solution set. Based on this index, the URL is recommended in such a way that the URL is more relevant and Threat is minimized. The architecture's performance is calculated and compared with the baseline approaches and it is clearly observed that the proposed trust-based URL recommendation system is dominating in terms of performance and attained a precision and accuracy of 96.84% and 95.87% respectively.