Web Log Clustering Analysis And Applications

With the rapid development of internet technology,people can get more information .But we find that getting the information which we indeed need is very hard,in other words, we lose our way in the cyberspace.Data Mining,which means getting knowledge from the database at the beginning.Now it is used in the cyberspace.We can get the knowledge from the cyberspace with the use of data mining technology. We discuss the definitions ,the roles.the methods of data mining in details.KD from web log is one way of using this technology.People pay more attention to this field mainly because web log has a very strict format. We can benefit from this character of it.First we discuss the clustering technology in depth.The model of clustering ,the method and how to evaluate the clustering are all included.We bring forward two definitions including dirty data and noise data, in the pretreatment process.We can get a more clear model with this two definitons.All the knowledge should be used in practice.Page recommendation is our way using the knowledge we get from the log.During this process,we bring forward a new score function which is used in the prediction evaluation.We consider more parameters in our function,in that way we can use the knowledge mining from the log more efficiently.At last,we design a new model namely LogDiscover.All our new ideas are used in this model.With the data from the model,we find these ideas are realistic and valid.