The increasing availability of different kinds of data and the spiralling availability of modelling techniques coming from statistical/machine learning have opened the way to new insights about the economy and a wider information set for policymakers.
Banca d'Italia will provide a two-day webinar to illustrate recent advances in economic modelling, harnessing big data along with machine learning approaches.
The classes will cover a wide range of topics including:
Theory:
- Machine learning (Deep Learning) for economic forecasting (e.g. Random forests, bagging and bootstrapping techniques, etc.);
- Regression with regularization techniques (modelling and estimation with many covariates or strong non-linearities);
- Text mining for article classification and sentiment analysis;
Empirics:
- Large granular structured or unstructured data sources (administrative data, web data, news and blogging platforms, payment data, textual data);
- Distributed computing for machine learning problems and data cleansing (matching, filtering or cleaning techniques, SQL-like techniques, regular expressions);
- Big data extractions for topics relating to firms, households, finance, labour markets, or government (e.g. Google data, data from Twitter, news from Factiva, etc.)
The event will take place from 25 to 26 February.
Participation is by invitation only and is reserved to central banks of selected new EU Member States, EU candidates and potential candidates, countries included in the EU's neighbourhood policy and other emerging economies.
