Chapter 8 - Trust discovery and information retrieval using artificial intelligence tools from multiple conflicting sources of web cloud computing and e-commerce users

Abstract

Cloud server adoption turns out to be increasingly well known and profoundly requested because of the unlimited information contribution and retrieval from anyplace and anytime. The existing method has lack of adaptability issues with a trust value calculation based on multidimensional cloud service providers. The truth content discovery algorithm is designed to produce trustworthy information in minimal time along with various types of media information. The technique provides trustworthy information from various news sources with minimal prediction error. It reduces the retrieval time of the query and ranks the news content index based on query matching. Based on an experimental evaluation, the proposed TCD+J48 method provides better results than existing methodologies. It minimizes the query retrieval time (QRT) by 0.9275s, mean absolute errors (MAE) by 1.22 s and roots mean squared error (RMSE) by 5.14 s. The trustworthy and scalable service providers algorithm (TSSP) is proposed for analyzing the design and relationship among the clients, the broker, and cloud service providers. The trust-based method effectively reduces the cloud client burden and improves the system stability. The proposed technique works based on the data entropy hypothesis to assess the multiattribute-based decision-making. Based on an experimental test, this method reduces the system execution time by 2msec, the communication cost by 9.33%, and improves the trust scores by 39.33% compared to existing methodologies in the respective parameters and tasks. An efficient feature extraction and classification (EFEC) algorithm is also proposed to extract a feature from opinion words. An EFEC algorithm is used to predict the quantity of positive and negative review opinions. According to an experimental assessment, the proposed technique improves accuracy by 15.05%, precision by 13.7%, recall by 15.59%, and F-measure by 15.07% of the proposed framework compared to existing techniques.