Abstract
Our main motivation is the data access model and aggregation algorithm for middleware by R. Fagin, A.
Lotem and M. Naor.
They assume data attributes in a variety of repositories ordered by a grade of attribute values of objects. Moreover they assume the user has an aggregation function, which eventually qualifies an object to top-k answers.
In this paper we adopt a model of various users (there is no single ordering of objects in repositories and no single aggregation) with user preference learning algorithm on the middleware side. We present a new model of repository for simultaneous access by many users.
The model is an extension of original model of Fagin, Lotem, Naor. Our solution is based on a model of fast learning of user preferences from his/her reactions.
Experiments are focused on the performance of top-k algorithms (both TA and NRA) using data integration on an experimental prototype of our solution. Cache size, network latency and batch size were the features studied in experiments.