Seminario didattico: Integration of Data Sources: Statistical Matching and other related methods

  • Data: 09 ottobre 2017 dalle 14:00 alle 16:00

  • Luogo: Dipartimento di Scienze Statistiche - v.le Berti Pichat - Aula VII

Relatore
Marcello D’Orazio
ISTAT e FAO

Sommario
Data integration techniques permit to investigate complex phenomena by exploiting sources of data already available. The methods can be classified according to their goal. Record linkage techniques aim at identifying the same unit in different data sources through key variables (name, address, etc.). On the contrary, the goal of statistical matching (data fusion) techniques consists in exploring relationship between variables not jointly observed in a single data source. Data fusion techniques are commonly applied in marketing or social research areas to integrate data collected through independent sample surveys referred to the same target population.

This seminar will provide insights on statistical matching, highlighting main areas of application and the frequently encountered problems. The seminar will also cover recent proposals of applying statistical matching techniques to integrate a sample survey with a larger data source not arising from a sample survey.

Organizzazione
Daniela Cocchi