vendredi 19 août 2016

cluster analysis for similar user navigation behavior

i have processes a log file for one web site contain 127 Web pages, and applying data cleaning and user identification process, instead of applying frequent item set or Apriori algorithm i would like to discover similar users based on browsing behavior towards these 127 Web pages?

    time stamp  Page-visit
user1   time    A1
user1   time    A2
user2   time    A3
user2   time    A1
user3   time    A4
user1   time    A26
user1   time    A100
user2   time    A1
user1   time    A14
user4   time    A88

i have browsing some methods for cluster analysis in log file i have build the following matrix for my users:

User Page View matrix

So, i think this matrix build over all log file .. my log file period one month so when consider this matrix how about frequency variables for each columns.. its required to converted to some other values?

is the procedure correct for above matrix when calculated over one month for all users?

How do we got similar users and what is similarity measure can be used for this kind of issue?

Aucun commentaire:

Enregistrer un commentaire