Adaptivní personalizovaná navigace řízená metadaty

Abstract

The work presents a scheme for navigating user based on metadata. We employ concept space that describes the problem domain by concepts and relations between them. Algorithm for concepts space evaluation was developed and is used to calculate an evaluation of a concept according its position and evaluation of its related concepts. Document which are presented to the users have defined a set of concepts which desribes them. For user an achived knowledge set is maintained where user's actual knowledge is stored in a form of known concepts. Navigation scheme takes the user's achived knowledge set and concepts sets of documents and chooses such documents which will be best beneficial for users in meaning of gaining new knowledge. For choosing the documents we employ numerous metrics which implement diferent approaches in learning of new knowledge such as choosing documents where user knows the most of the concepts and others or choosing document according concepts similarity to presented document. The chosen documents are presented to the user in a form of menu where the best suitable document in on the top. The user's behavior and reactions to the presented documents help to shape the navigation prepared for them.