A semantically-enabled trust model for collaborative environments
The World Wide Web has revolutionized the way we find, exchange, reuse and integrate information at a rate inconceivable only a generation ago; but for all its perks it has one big weakness: trusting that the information retrieved is accurate. Web technologies and applications allows anyone to share information. This work aims to assist humans (and eventually autonomous agents) in the task of trusting Web resources, a very important step when reusing information found on the Web. In this research, we propose a new algorithm to tackle the challenge of assigning trust to Web resources. Our algorithm is inspired on Google's famous algorithm PageRank and generates values of trustworthiness of Web resources based on dynamically adjusted user credentials on different areas of study and history of interactions. The proposed algorithm was implemented in the Virtual Learning Commons, a bookmarking system, developed and hosted at the Cyber-ShARE Center of Excellence. The proposed algorithm was evaluated in an academic environment but can be used in other applications. The evaluation shows that the algorithm produced outperforms two out of the three algorithms we benchmarked against, while matching the results of the third. The evaluation results also show that the algorithm provides adequate trust assessments for Web resources. Future work includes the extension of this algorithm to consider additional variables (i.e., beyond user credentials) and a more comprehensive evaluation settings, including a more diverse subject population. ^ Efforts that facilitate Web users to assign values of trustworthiness to Web resources will contribute to the overall utility of the world wide web, it will leverage the knowledge of the world and will formalize said knowledge so that machine agents may use to our benefit.^
Reyna Cruz, Joaquin Andres, "A semantically-enabled trust model for collaborative environments" (2016). ETD Collection for University of Texas, El Paso. AAI10118792.