lp_zgl output

[erin00...@gmail.com]

Hi,

This is the output of lp_zgl (see below).

<node_name>TAB[<gold_label> <gold_score>]+TAB[<seed_label> <seed_score>]+TAB[<estimated_label> <estimated_score>]+TAB<is_test_node>TAB<node_MRR>

...

N881 L2 1.0 L2 1.0 L2 1.0 true 1.0

N883 L2 1.0       L1 1.0 L1 1.0 L1 1.0 true 1.0

N885 L2 1.0 L2 1.0 L2 1.0 true 1.0

N2417 L2 0.8366457772307115 L1 0.16335422276928854 false 0.0

N2416 L2 0.8483981615059183 L1 0.15160183849408174 false 0.0

N880 L2 1.0 L2 1.0 L2 1.0 false 1.0

...

My question is, why for some of the *test* nodes there are different number of fields?

(btw, adsorption and MAD algorithms on the same dataset do not produce this type of output)

Thank you!!

[Partha Pratim Talukdar]

I am not sure what you mean. Are you referring to the second column, i.e., "L2 1.0" in the rows with is_test_node == true? If so, then those are [<gold_label> <gold_score>]+

Partha

[Erin Skylar]

For N883, L2 1.0 is [<gold_label> <gold_score>], L2 1.0 is [<seed_label> <seed_score>], and the next two are both estimation pairs, L1 1.0 L1 1.0? 

In the training set, I only have one gold label for N883, and that is L2 with score 1.0.

--
You received this message because you are subscribed to a topic in the Google Groups "The Junto Label Propagation Toolkit Open Discussion" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/junto-open/BfVAt37fU4E/unsubscribe.
To unsubscribe from this group and all its topics, send an email to junto-open+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Partha Pratim Talukdar]

On Sun, Dec 14, 2014 at 10:15 PM, Erin Skylar <erin00...@gmail.com> wrote:

For N883, L2 1.0 is [<gold_label> <gold_score>], L2 1.0 is [<seed_label> <seed_score>], and the next two are both estimation pairs, L1 1.0 L1 1.0? 

Each line should have 6 fields in total (tab delimited). Are you sure the line for N883 is junto generated, as only that line seems to have more than 6 fields? Also, if it is, then did you happen to change the code?

Partha