Research And Application Of A New Self-organizing Features Algorithm Based On Gwrn

Web server log records the interaction between user and server information.And users' activity on the site implicts their needs and interests.If we analysis the Web Log,this can help to optimize the sites's organizational structer and improve web server performance and identifize potential custormers of electronic commerce and enhance the user's personalized quality of service.Self-organizing feature map network cluster the input data automatically.It has good adaptive ability and robustness,is widely used in data mining.However,SOM has its limitations.The ability to grow extra nodes is a potentially useful facility for a self-organising neural network. A network that can add nodes into its map space can approximate the input space more accurately, and often more parsimoniously, than a network with predefined structure and size, such as the Self-Organising Map.Based on the advantages of self-organizing neural networks in pattern clustering,we combine it with Growing When Required network(GWRN) and Hebb learning rule.A new self-organizing neural network that has three-layer structer is developed in this paper.This algorithm maintains good self-growth and self-organizing characteris of GWRN.The learning algorithm can add nodes whenever the network in its current state does not sufficiently match the input. In this way the network grows very quickly when new data is presented, but stops growing once the network has matched the data. The learning process is divided into competition stage and self-motivation stage.GWRN algorithm is used in competition stage,the network grows dynamic.Hebb learning rule is used in self-motivation stage.This will classifies the neurons that motive each other as a class.Finally, we take simulation tests and practical application to the algorithm that proposed in this paper.And compare it with GWRN algorithm.The main task in the simulation tests is choosing the best values of several parameters. Then apply the algorithm to real Web log data.Through analysising user access patterns,we can mining user's multi-interest and potential interest,so as to provide users with better personal web service.