作者: JULIANA YIM , HEATHER MITCHELL
DOI: 10.1111/J.1467-8454.2007.00326.X
关键词: Warning system 、 Statistical model 、 Artificial neural network 、 Actuarial science 、 Event (computing) 、 Financial distress 、 Computer science 、 Financial services
摘要: This paper looks at the ability of a relatively new technique, non-linear extension Granger thick model concept, hybrid ANN's, to predict failure financial service firms in Australia. These models are compared with traditional statistical techniques and conventional ANN models. The results suggest that neural networks outperform all other predicting for up two years prior event. suggests researchers, policymakers others interested early warning systems, network may be useful tool firm failure.