1. Self-Organized Web Usage Regularities

2. Problems of foraging information on WWW Slow accession Difficulty in finding useful information is related to balkanization of Web structure Difficulty to solve this fragmentation problem by designing an effective classification scheme. Solution: To seek Web regularities in user behavior.

3. Issues How to characterize the strong regularities in Web surfing in terms of user navigation strategies. How to present an information foraging agent-based approach to describing user behavior. Issues on web self-organization.

4. Related research works Web mining for pattern-oriented adaptation: to identify the inter-relationships among different websites, either based on the analysis of the contents in Web pages or based on the discovery of the access patterns from Web log files. Web data mining: computing association rules, detecting sequential patterns, and discovering classification rules and data clusters. Web usage mining: analysis of Web usage patterns, such as user access statistical properties, association rules and sequential patterns in user sessions, user classification and web page clusters based on user behavior.

5. Information Foraging Agent Model (IFAM) Objectives: to find the inter-relationship between the statistical observations on Web navigation regularities and foraging behavior patterns of individual agents.

6. IFAM – Artificial Web Space Artificial Web Space: a collection of websites connected by hyperlinks. D(c i,c j )=( Σ M k=1 (cw i k – cw j k ) 2 ) 1/2 Where D(c i,c j ) denotes the Euclidean distance between the content vectors of nodes i and j.

7. IFAM – Artificial Web Space Content Distribution Models T + |X c |, if i=j cw n i = |X c |, otherwise fx c ~ normal (0,   ) T ~ normal(  t,  t ) Where fx c: probability distribution of weight x c normal (0,   ): normal distribution with mean 0 and variance  . T: content (increment) offset on a topic  t: mean of normally distributed offset T  t: variance of normally distributed offset T

8. IFAM – Artificial Web Space Power-law distribution: T + |X c |, if i=j cw n i = |X c |, otherwise fx c ~   (X c + 1) – (   +1), X c > 0,   > 0 Where fx c: probability distribution of weight X c   : shape parameter of a power-law distribution T: content (increment) offset on a topic

9. IFAM – Artificial Web Space Constructing an Artificial Web 1. For each topic i 2.Create node content vectors End 3.For each node i 4. Initialize the link list of node i 5. For each node j 6.If D(c i,c j ) < r 7.Add node j to the link list of node i 8.Add D(c i,c j ) to the link list of node i end

10. IFAM – Foraging Agents Interest Profiles p m = [ pw m 1, pw m 2 …pw m i …pw m M ] pw m i P mi =  m j=1 pw m i H m = -  m j=1 p mi log(p mi ) Where p m: preference vector of agent m pw m i : weight of preference on topic i H m: interest entropy of user m

11. IFAM – Foraging Agents Interest Distribution Models 1. Normal distribution: pw m i = X p fxp ~ normal(0,  u ) Where normal(0,  u ) denotes the normal distribution with mean 0 and variance  u 2. Power-law distribution: pw m i = X p fxp ~  u (X p +1) -  u + 1, X p > 0,  u > 0 Where  u denotes the shape parameter of a power- law distribution

12. IFAM – Foraging in an Artificial Web Space Random agents: have no strong interests in any specific topics. Rational agents: have specific interested topics in mind and they forage in order to locate the pages that contain information on those topics. Recurrent agents: Recurrent agents are those who are familiar with the Web structure and know the whereabouts of interesting contents.

13. IFAM – Foraging in an Artificial Web Space Agent Preference Updating: depending on how much information on interesting topics the agent has found and how much the agent has absorbed such information. P m (  ) = P m (  - 1) - c n pw m i = 0, for pw m i (  ) < 0, i = 1…M Where denotes an absorbing factor in [0,1] that implies how much information is accepted by agents on average. P m (  ) and P m (  - 1) denote an agent’s preference vector after and before accessing information in page n, respectively.

14. IFAM – Foraging in an Artificial Web Space Motivation Functions f log (  m t v ) ~ normal(  m,  m ) Where  m and  m denote the mean and variance of the log-normal distribution of  m t v, respectively  m t v =  m e  m step Where  m and  m denote the coefficient and rate of an exponential function. Step denotes the number of pages/notes that an agent has continuously visited.

15. IFAM – Foraging in an Artificial Web Space Rewarding Function  R t =  M i=1 (pw m i (  -1) - pw m i (  )

16. IFAM – Foraging in an Artificial Web Space Foraging 1. Initialize the nodes and links in an artificial Web space 2.Initialize information foraging agents and their interest profiles 3.For each agent m 4. While the support for the agent S min_support m 5.Find the hyperlinks inside node n that the agent is presently in 6. Select, based on pk, the hyperlink that connects to the next-level page 7.Forage to the selected page 8.Update the preference weights in the agent’s interest profile 9.Update the support function of the agent End 10. If the support for the agent S > max_support m 11.Agent m is satisfied with the contents and leaves the Web space Else 12.Agent m is dissatisfied and leaves the Web space End

17. Result The experiment shows that by applying a weighted linear-regression method, the higher the occurrence rate of a depth or a link-click-frequency is, the higher the weight will be.

18. Self-Organized Agent To support adaptive organizations between agent, adding modeling technique allows agents to model their interactions with the environment and to recognize and manipulate new environmental scenarios to achieve organizational goals.