Naslov
Using Behavioral Statistical Scorecards in Portfolio Management and Business Planning

Autori
Klepac, Goran ; Kliček, Božidar

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
APPLIED STATISTICS / INTERNATIONAL Conference Applied Statistics (2007 ; Ribno) Program and abstracts / International Conference Applied Statistics 2007, September 23-26, 2007, Ribno (Bled), Slovenia ; [edited by Gaj Vidmar and Janez Stare]. - Ljubljana : Statistical Society of Slovenia, 2007 - : INTERNATIONAL Conference Applied Statistics (2007 ; Ribno) Program and abstracts / International Conference Applied Statistics 2007, September 23-26, 2007, Ribno (Bled), Slovenia ; [edited by Gaj Vidmar and Janez Stare]. - Ljubljana : Statistical Society of Slovenia, 2007, 2007, 65-65

Skup
Applied Statistics

Mjesto i datum
Ribno, Slovenija, 23.09.2007. - 26.09.2007

Vrsta sudjelovanja
Poster

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
: behavioral statistical scorecards, data mining, Bayesian networks

Sažetak
Behavioral statistical scorecards have increased their popularity with BASELII standards in financial institutions. This article recognizes new potential and fruitful application area of behavioral statistical scorecards for portfolio management and holistic business planning from the perspective of risk measurement. This methodology could not be applied only in financial institutions but also in different types of business like insurance or trade. Risk measurement through behavioral statistical scorecards can be the base for business planning with the perspective of client/customer future behavior considering delays in paying invoices, delays in paying loans (default), fraud, churn and similar risky behavior. Behavioral statistical scorecards method can clearly explain influences on target variable and perform client/buyer profiling. Campaign planning could be focused on potential clients with low risk profile from the perspective of fraud, delays in paying or/and churn, depending on company policy (conservative/liberal). All those parameters could be measured periodically with trend monitoring of risk parameters on whole portfolio or in portfolio segments. Behavioral statistical scorecards method can be useful for the construction of an early warning system that is sensitive to structural changes in portfolio. Another potential application is "what-if" or sensitivity analysis with Bayesian networks using calculated risk measures (scores from statistical scorecard) as an input. This article describes several innovative applications performing problem analysis first, then explaining the underlying theoretical model, and then describing the practical examples of application. Theoretical and implementation issues of these applications are discussed, and future research directions are outlined

Izvorni jezik
Engleski

Znanstvena područja
Informacijske i komunikacijske znanosti

POVEZANOST RADA