Application of Data Mining Methods for Pattern Recognition in Negotiation Support Systems

Abstract

Data mining methods have long been used to support organisational decision making by analysing organisational data from large databases. The present paper follows this tradition by discussing two different data mining techniques that are being implemented for pattern recognition in Negotiation Support Systems (NSSs), thereby providing process assistance to human negotiators. To this end, data from several international negotiation experiments via NSS Negoisst is used. Consequently, a suitable data representation of the underlying utility data and communication data has to be created for the applicability of data mining. Each generated data type needs individual processing treatments and almost all data mining methods lose their feasibility without a correct data representation as consequence. Once a correct data representation is found, the potential for pattern recognition in electronic negotiation data can be evaluated using descriptive and predictive methods. Whilst Association Rule Discovery is used as a descriptive technique to generate essential sets of strategic association patterns, the Decision Tree is applied as a supervised learning technique for the prediction of classification patterns. The extent to which reliable as well as valuable patterns can be derived from the electronic negotiation data and valuable predictions can be generated is examined in this paper.