search query: @indexterm Cluster analysis / total: 167
reference: 87 / 167
| Author: | Oestreicher, A. Piotrowski-Allert, S. |
| Title: | Klassifikation und Beurteilung von Unternehmen mit Hilfe von Selbstorganisierenden Neuronalen Netzen |
| Journal: | Schmalenbachs Zeitschrift für Betriebswirtschaftliche Forschung
1996 : VOL. 48:4, p. 335-371 |
| Index terms: | FINANCIAL STATEMENTS FINANCIAL ANALYSIS NEURAL NETWORKS CORPORATE FINANCIAL MODELS CLUSTER ANALYSIS CLASSIFICATION AND CODING |
| Language: | ger |
| Abstract: | This article presents a model for analyzing financial statements with self-organizing neural networks. The model is based on an algorithm developed by Kohohen which transforms similarities between input data into space coordinates. The input data consists of 22 features which are derived from the 105 financial statements and annual reports examined. It is shown that similarities between companies can be transformed into coordinates on an two- dimensional map. On the basis of these coordinates a cluster analysis leads to the formation of company classes. The feature map does not give any information about hierarchies existing within and between the groups. In our case this analysis results in seven features determing the classification. In order to classify new data the results from the Kohonen-model can be used. |
SCIMA