The Sensitivity of machine learning techniques to variations in sample size: a comparative analysis

Javier de Andrés Suárez ^[1] ; Pedro Lorca Fernández ^[1] ; Elías Fernández Combarro Álvarez ^[1]
1. [1] Universidad de Oviedo
  
  Universidad de Oviedo
  
  Oviedo, España
Localización: The International Journal of Digital Accounting Research, ISSN 1577-8517, Vol. 2, Nº. 4, 2002, págs. 131-155
Idioma: inglés
DOI: 10.4192/1577-8517-v2_5
Enlaces
- Texto completo

Resumen
- A comparative analysis of the performance of some well-known classification techniques (Discriminant Analysis, Quinlan's See5, and Neural Networks) and certain machine learning systems of recent development (ARNI, FAN and SVM) is conducted. The chosen classification task is the forecasting of the level of efficiency of Spanish commercial and industrial companies. Assignment of the firms is made upon the basis of a set of financial ratios, which make a high dimension feature space with low separability degree. In the present research the effects on the accuracy of variations of each technique in the estimation sample size are measured. The main results suggest that ARNI and See5 yield the best results, even with small sample sizes.