Naslov
Descriptive modeling of systemic banking crises

Autori
Gamberger, Dragan ; Lučanin, Dražen ; Šmuc, Tomislav

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Discovery Science 15th International Conference, DS 2012 / Ganascia, J-G ; Lenca, P. ; Petit, J-M - Heidelberg : Springer, 2012, 67-80

Skup
Discovery Science 15th International Conference, DS 2012

Mjesto i datum
Lion-sur-Mer, Francuska, 29.10.2012. - 31.10.2012

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Subgroup discovery; Knowledge discovery; Banking crises

Sažetak
The topic of the work is detection of connections between occurrences of systemic banking crises and values of socio-economic indicators in time frames of three years before outburst of crises. For this task we have used the list of banking crises in the period 1976-2007 prepared by the International Monetary Fund that we have connected with publically available Word Bank data. For the analysis a machine learning methodology based on subgroup discovery has been used. The main result is that demographic indicators have been detected as most relevant. At first place this is the indicator of percentage of total population that is in the age group 15-64 years. This indicator is present in both developed models and presents a condition related to a high number of crises outbursts. In the first model this indicator is connected with the indicator of annual percentage of money and quasi money growth while in the second model it is connected with the indicator of life expectancy for male population. For the analysis especially interesting is the second model because it includes decreasing or very low positive trend in active population in a period before the onset of the crises. The importance of the result is in the fact that such situations may be expected in the near future in many developed and developing economies.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo

POVEZANOST RADA