User-based document ranking
Mario Kubek, Phayung Meesad, Herwig Unger · Proceedings of the 3rd International Conference on Communication and Information Processing · 2017
The ranking of web documents is the key to present really significant results of a keyword-based search at first to users. Standard methods do not consider any user feedback so far and usually need a tremendous computation performance. The presented new approach considers access to web documents as the sole evaluation criterion and ensures that frequently accessed pages are recognised, can be found fast and are presented at the beginning of a search session. Furthermore, its locally working algorithms have been conceived to be applied in fully decentralised search systems in which no global knowledge base is available for this purpose.