Mining the Attitudes of Conflicting Human Agents

B. Galitsky and B. Kovalerchuk (USA)


Behavior of human agents, multi-agent conflict, commu nicative actions, machine learning, customer complaints.


We build the generic methodology based on machine learning and reasoning to detect specific attitudes of hu man agents. This methodology is applied to textual as well as structured data on inter-human conflicts. The pro posed approach is applicable to a wide range of domains where mining for the attitudes of involved agents is cru cial. Human attitudes are determined in terms of commu nicative actions of agents; machine learning is used be cause it is rather hard to identify attitudes in a rule-based form directly. One of the main problems to be solved while assisting inter-human conflict resolution is how to reuse the previ ous experience with similar agents. A machine learning technique for handling scenarios of interaction between conflicting human agents is proposed. The developed scenario representation and comparative analysis tech niques are applied to the classification of textual customer complaints. Then we employ the scenario knowledge rep resentation technique in such problems as predicting an outcome of international conflicts, and mining emails for suspicious emotional profiles. Successful use of the pro posed methodology in rather distinct domains shows its adequacy for mining human attitude-related data in a wide range of applications.

Important Links:

Go Back