Research On Distributed Data Mining System Based On Dynamic Agent Technology

Distributed system is widely used because of its ease of use, convenience and highperformance.However, in a distributed system of rapid growth, These problems in acertain extent play a restraining effect on the promotion of distributed systems.Thispaper adoptsdynamically Agent and technology in data mining process that theproblems encountered is based on a distributed systemFirstly,use the hierarchical system structure to make the system flexible andexecute the Agent data mining results.Secondly, the use of special control agent tomanage the local data warehouse, data mining tasks simultaneously the receptioncenters Agent released, and the breakdown of the task and data mining summary of theresults and inspection; Finally, the executive Agent specific data mining tasks.And theresulting knowledge is fed back to the control Agent.Hierarchical system structure has astrong flexible line and scalability.And through center Agent task allocation controlachieve load balancing.Secondly, through the packet achieves communication between agents, through thecommunication message format of design specifications,educes the amount of datatransmission in the network, but also reduces the degree of coupling between eachAgent.Data inconsistency distributed data mining is enhanced.Finally,ant colony algorithm is improved by dichotomy, so that the improved antcolony algorithm can be better to adapt to the distributed environment.In DynamicAgent Distributed Data Mining system will be studied by improved ant colonyalgorithm.Dynamic Agent-based distributed data mining system which is studied in thispaper,adopts the hierarchical system architecture to improve the flexibility andscalability of the system,and reduce the amount of data transmission in the networkthrough standardized communication format,and make improvements through the antcolony algorithm, Data mining algorithms of Executive Agent will be studied.