Mining ‘Living’ Data - Providing Context Information to a Negotiation Process

abstract

Negotiation is the process whereby two (or more) individual agents with conflicting interests reach a mutually beneficial agreement on a set of issues. In negotiation the exchange of information is as important as the exchange of offers. During a negotiation, an agent may actively acquire contextual information that it may, or may not, choose to place on the negotiation table. This work presents a multi-agent framework for negotiation support and discusses on more details a framework for online data mining that supports the identification of contextual information and providing it on negotiator’s demand. The smart data mining agents that support the negotiation agent, are expected to operate under time-constraints and over dynamically changing corpus of information. They need to determine the sources of information, the confidence and validity of these sources and a way of combining extracted information. All transactions, including complex requests for information and combination of results, are managed as business processes. There are a number of challenges that the smart data mining architecture is addressing, including critical pieces of information being held in different repositories; non-standard nomenclatures; radically different data types and models; possible duplicative, inconsistent and erroneous data; and possible high rate of change of the models representing data content. The mining and discovery procedures include: (i) mining the opponent’s profile information (this is a broad group of methodologies in which adapt and further develop: user-centric and site-centric data mining methods, methods for mining social