Algorithms For Dummies

  Getting the Right Web page 

     215

there are no arrows, but the direction of an edge is represented by a thicker line 

entering a node. For example, a surfer can go from node 4 to node 6 because there 

is a thick line entering node 6 from node 4. However, the surfer can’t go from 

node 6 to node 4 because the line entering node 4 from node 6 is thin.

The second graph isn’t strongly connected. It presents a trap for a random surfer 

because the second node has no outbound links, and a user visiting the page could 

stop there and find no way out. This isn’t an unusual event considering the struc-

ture of the web, but it could also show a spammer artifact, such that the spammer 

created a spam factory with many links that direct to a dead end site in order to 

trap web surfers. Figure 11-2 shows the output of the following code, which was 

used to display this graph.

np.random.seed(2)

pos=nx.shell_layout(Graph_B)

nx.draw(Graph_B, pos, arrows=True, with_labels=True)

plt.show()

Another situation that may be natural or the result of a spammer’s action is a 

spider trap. It’s another dead end for a surfer, this time not on a single page but 

on a closed site that lacks links to an outside network of pages. Figure 11-3 shows 

the output of the following code, which was used to display this graph.

np.random.seed(2)

pos=nx.shell_layout(Graph_C)

nx.draw(Graph_C, pos, arrows=True, with_labels=True)

plt.show()

FIGURE 11-2: 

A dead end.

---

216

 

   

  PART 3