Foto del docente

Marco Patella

Associate Professor

Department of Computer Science and Engineering

Academic discipline: ING-INF/05 Information Processing Systems

Research

Keywords: Multimedia databases BigData analytics Metric trees Preference queries

  • Similarity searching: the goal of this activity is the development of efficient techniques for the processing of similarity queries.
  • Preference-based searching: the focus is the definition of efficient algorithms for processing queries based on "qualitative preferences".
  • BigData analytics: this activity has the goal of creating effective and efficient solutions for query processing in the BigData context.


  • Similarity searching: similarity queries allow the user to search for database objects that are similar, according to a specific similarity criterion, to a given query object. In order to allow an efficient processing of such queries, the similarity criterion is usually based on the definition of a distance function that satisfies the metric postulates. The goal of this activity is the development of efficient techniques for the processing of similarity queries employing such properties of the distance function in order to avoid a costly sequential scan of the whole database.
  • Preference-based searching: the focus is the definition of efficient algorithms for processing queries based on "qualitative preferences". Qualitative preferences allow for a higher flexibility in the formulation of queries based on different criteria, since they do not force the user to specify a single sorting criterion to evaluate relevance of data wrt the query. Instead, they allow to establish a simple preference relation among data. Such flexibility, however, enforce additional requirements for the specification of query processing algorithms.
  • BigData analytics: the goal of this activity is that of extending techniques for efficient processing of both similarity and preference-based queries (that are the focus of previous research activities) to the distributed scenario and, in particular to BigData systems.

Latest news

At the moment no news are available.