You just nailed the problem! Please consider the following arbitrary network and the PageRank scores below:
[image: Inline image 2] [image: Inline image 1] As you see, both A and B user nodes contribute differently to the network. They interact with K_i products and express certain sentiments (just one sentiment in this example for simplicity: H). Yet they have the exact PageRank because they both don't receive any voting from other nodes (users receiving votes by products and sentiments doesn't make sense in this case). I am thinking that there should be anther way to quantify the interaction given the different activities. Perhaps using the multilayer network approach will reveal more info. If you think of something else that can hight such activities please feel free to share. Thanks very much, -Ahmed On Mon, Mar 21, 2016 at 5:34 AM, Tamas Nepusz <[email protected]> wrote: > Hi, > > Multiple nodes having the exact same PageRank values are not > necessarily a cause for concern - there could be logical explanations > for it. For instance, any "source" vertex in a directed graph has the > exact same PageRank value due to the fact that these vertices have no > inbound edges, so the only way the random walk can reach them is via > teleportation - and teleportation to a certain node has a well-defined > probability that comes straight from the damping parameter and the > number of nodes in the graph. > > T. > > > On Sun, Mar 20, 2016 at 6:22 PM, Ahmed Abdeen Hamed > <[email protected]> wrote: > > Thanks, this is helpful! > > > > I am experiencing a phenomenon where certain node types in my network > have > > the exact PageRank value. A while back, I recall that there was an issue > > with PageRank, but you helped me to get around it. So, I am not exactly > sure > > if this current issue is related to the previous PageRank issue or > perhaps > > it is related to the data itself. Please note, this issue is only > > demonstrated with one particular type of nodes. So, in the same network > of > > different node types, all nodes have different values except that type. I > > tried this with different networks with a similar structure but still > > getting the exact same results. If this is indeed a data issue, I am > > thinking of using a multilayered network approach to identify the > > significant nodes spanning across multiple layers. Let me know if you > have > > any insights and I apologize if this is this is too vague! > > > > Thanks very much, > > > > -Ahmed > > > > On Sat, Mar 19, 2016 at 5:03 PM, Tamas Nepusz <[email protected]> > wrote: > >> > >> Hi, > >> > >> It depends on what you want to do with it. The easiest is to treat it > >> as a single network and add a vertex attribute that describes which > >> layer a given vertex belongs to. > >> T. > >> > >> > >> On Sat, Mar 19, 2016 at 7:04 PM, Ahmed Abdeen Hamed > >> <[email protected]> wrote: > >> > Hello friends, > >> > > >> > Is there a way to represent a multilayered network using igraph? > >> > > >> > Thanks very much, > >> > > >> > -Ahmed > >> > > >> > _______________________________________________ > >> > igraph-help mailing list > >> > [email protected] > >> > https://lists.nongnu.org/mailman/listinfo/igraph-help > >> > > > > > >
_______________________________________________ igraph-help mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/igraph-help
