Machine Learning for Negotiation Knowledge Discovery in e-Marketplaces

Abstract

The level of autonomy and the efficiency of e- Marketplaces can be improved if automated negotiation support is available. Some parametric learning negotiation models have been proposed recently. These models allow a negotiator to learn the opponents' preferences based on previous offer exchanges. Nevertheless, these models make strong assumptions about the particular negotiation mechanism employed by the respective negotiation agent. This paper illustrates the design, development, and evaluation of a non-parametric negotiation knowledge discovery method which is underpinned by the well-known Bayesian learning paradigm. This method can discovery vital information about a negotiator's preferences without making any assumption about the underlying negotiation mechanism employed by the negotiator. According to our empirical testing, the proposed negotiation knowledge discovery method can speed up the negotiation process while maintaining the negotiation effectiveness. Our research work opens the door to the development of intelligent negotiation mechanisms to enhance modern e-Marketplaces.