The information retrieval model is the core of the information retrieval. Nowadays, with the recent advances in the theory and practice of concept lattice, many researchers are using concept lattice as a basis for constructing information retrieval model, the representative of which is the concept lattice based ranking model (CLR). However, the CLR is only based on the shortest path linking the query to the documents. It is not well defined. To resolve the existing problem, first, this paper presents a concept lattice based user model to represent user interest in concept lattice structure. A novel concept lattice based user modeling algorithm is also provided. Second, this paper proposes a concept lattice based personalized ranking model (CLPR), which measures the similarity among query, user interest model and documents according to the relation between query and user interest based on concept lattice. The similarity computation of CLPR contains more information than traditional information retrieval models, and moreover, the CLPR can identify the new interest of users and find the documents of new interest. Our experiments show that documents retrieved by CLPR achieve steadily higher measures of precision than the CLR and the traditional vector space model. And CLPR is feasible to common search engine users.

The performance of a controller is usually characterized by several attribute indexes. In this context, this paper applies multi-attribute decision-making techniques to the assessment of control performance, which takes into account both stochastic and deterministic performance indexes. Case studies demonstrate the validity of the proposed method.

No abstract