Naslov
Financial ratios clustering methodology

Autori
Župan, Mario ; Budimir, Verica ; Letinić, Svjetlana

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

Izvornik
2nd International Conference on Computational Finance / Guerra, Manuel ; Janela, João - Lisabon, 2017, 211-216

Skup
2nd International Conference on Computational Finance

Mjesto i datum
Lisabon, Portugal, 04.09.2017. - 08.09.2017

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
financial ratios, liquidity, clustering, data mining

Sažetak
This paper explores the possibilities of clustering algorithm on a sample of 2, 200 small businesses and 3 financial ratios. The procedure is described in details: preprocessing phase manipulation with extreme values, application of two types of hierarchical clustering, two measures of distance, dimension reduction and visualisation by the self-organizing map (SOM) algorithm. The accuracy of the algorithms is proved visually and with the Tukey's honestly significant difference (HSD) post hoc test. Algorithms had not had a difficult task because artificial groups of enterprises were created by modifying real dataset. Several important facts about the procedure and methodology of clustering and visualisation of financial data have been established. Final output of this work is transparent and aggregated information about the financial position of the sample.

Izvorni jezik
Engleski

Znanstvena područja
Ekonomija

POVEZANOST RADA