Μέθοδοι σύστασης πολυμεσικών δεδομένων βασισμένες σε τεχνικές μηχανικής μάθησης
Λαμπρόπουλος, Αριστομένης Σ.
SubjectIntelligent agents (Computer software) ; Electronic data processing ; Multimedia systems ; Computer networks
Recent advances in electronic media and computer networks have allowed the creation of large and distributed repositories of information. However, the immediate availability of extensive resources for use by broad classes of computer users gives rise to new challenges in everyday life. These challenges arise from the fact that users cannot exploit available resources effectively when the amount of information requires prohibitively long user time spent on acquaintance with and comprehension of the information content. Thus, the risk of information overload of users imposes new requirements on the software systems that handle the information. Such systems are calls Recommender Systems (RS) and attempt to provide information in a way that will be most appropriate and valuable to its users and prevent them from being overwhelmed by huge amounts of information that, in the absence of RS, they should browse or examine. In this thesis, firstly, it explored the use of objective content-based features to model the individualized (subjective) perception of similarity between multimedia data. It presents a content-based RS which constructs music similarity perception models of its users by associating different similarity measures to different users. The results of the evaluation of the system verified the relation between subsets of objective features and individualized (music) similarity perception and exhibits significant improvement in individualized perceived similarity in subsequent recommended items. Secondly, it addressed the recommendation process as a hybrid combination of one-class classification with collaborative filtering.