Naslov
PREDICTING SUSTAINABLE FINANCIAL PERFORMANCE USING CASH FLOW RATIOS: A COMPARISON BETWEEN LDA AND ANN METHOD

Autori
Aljinović Barać, Željana

Izvornik
Sarajevo business and economics review (1986-5473) 31 (2011), 1; 33-58

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni

Ključne riječi
cash-flow ratios; discriminant analysis; neural networks; financial performance prediction

Sažetak
The aim of this paper is to provide empirical evidence concerning the ability of cash-flow ratios to predict sustainable company’s performance using two different methodological approaches: linear discriminant analysis, as the most popular statistical method and artificial neural networks as more sophisticate method. Cash flow ratios are used for sustainable financial performance prediction rather than accrual ratios, as they are less likely to be affected by possibilities of accounting manipulations. The results of our study indicate that cash flow ratios significantly differ between sustainable and non-sustainable financial performance companies. Furthermore, it is shown that both discriminant analysis and artificial neural networks models have relatively good accuracy and provide similar classification results. The best results are provided when discriminant analysis and neural networks methods were combined so only the ratios selected by discriminant analysis were used as in inputs in neural networks model. Contribution of this paper can be found in: (1) the comprehensive review of corporate performance or financial distress prediction models topic, (2) development of sustainable financial performance prediction models based on discriminant analysis and artificial neural networks methodology and (3) comparison of discriminant analysis and artificial neural networks models accuracy in short-term and medium-term prediction.

Izvorni jezik
Engleski

Znanstvena područja
Ekonomija

POVEZANOST RADA