DBDM - Database and Data Mining Group

The DBDM Group carries out its research activity in various areas within the field of data mining and databases. Data mining tackles the study of algorithms aimed at discovering "hidden" information stored in large data collections. The problem of extracting such information is a rather complex due to the large number of variables that must be taken into account. Moreover, the complexity significantly increases when increasing the volume of the data to be processed (Big Data). Big Data underlines the limits of the existing data mining techniques and poses new challenges for the design of novel algorithms to address data analysis. The research activity of the DBDM Group focuses on the study of algorithms for diverse data mining tasks on Big Data, including association rule mining to discover correlation among data at different abstraction levels, the extraction of knowledge for performing predictions (classification task), grouping of similar data (clustering task). The data analytic algorithms, in the Big Data context, must provide the necessary scalability, accessibility, extensibility, and flexibility. The proposed algorithms are validated in different application contexts (e.g., network traffic data analysis, text mining and social network applications, health and medical applications, financial applications).

Skills

  • Design on novel algorithms for Big Data analysis
  • Integration of data mining techniques in relational databases, e.g., by defining novel types of indices
  • Definition of disk-based indices to support data mining algorithms
  • Classification of structured (relational), semi-structured (XML) and unstructured data (text documents)
  • Text mining for social network analysis
  • Application of data mining techniques to network traffic data
  • Application of data mining techniques to clinical and biological data
  • Application of data mining techniques to sensor network data

Projects and publications

  • Selected recent publications