Specialized KD-Agent for Knowledge Ecosystems

Abstract

One of the base elements of any knowledge ecosystem is a software agent. The agent receives data about the internal events of the ecosystem, interprets data and executes commands that affect the environment. The paper proposes an option for the implementation of the specialized Knowledge Discovery agent (KD-agent). The input for the agent is the a priori dictionary of features and the training set. As the outcome of the agent activity previously unknown patterns are revealed and can be interpreted within the subject domain. The effectiveness of the proposed approach is demonstrated on the example of model data analysis.