TREC WT10g

Metadata

Code		`WT`
Internal name		`trec-wt10g`
Name		TREC WT10g
Data source		http://ir.dcs.gla.ac.uk/test_collections/access_to_data.html
Availability		Dataset is available for download
Consistency check		Dataset passed all tests
Category		Hyperlink network
Network format		Unipartite, directed
Edge type		Unweighted, no multiple edges
Reciprocal		Contains reciprocal edges
Directed cycles		Contains directed cycles
Loops		Does not contain loops

Size	n =	1,601,787
Volume	m =	8,063,026
Loop count	l =	0
Wedge count	s =	4,375,162,265
Claw count	z =	17,206,337,249,825
Cross count	x =	84,325,207,721,815,904
Triangle count	t =	21,071,515
Square count	q =	12,596,865,743
4-Tour count	T₄ =	118,288,933,500
Maximum degree	d_max =	25,663
Maximum outdegree	d⁺_max =	25,604
Maximum indegree	d⁻_max =	7,237
Average degree	d =	10.067 5
Fill	p =	3.142 60 × 10⁻⁶
Size of LCC	N =	1,458,316
Size of LSCC	N_s =	470,441
Relative size of LSCC	N^r_s =	0.293 698
Diameter	δ =	112
50-Percentile effective diameter	δ_0.5 =	7.911 05
90-Percentile effective diameter	δ_0.9 =	11.104 7
Median distance	δ_M =	8
Mean distance	δ_m =	8.703 68
Gini coefficient	G =	0.637 390
Balanced inequality ratio	P =	0.266 576
Outdegree balanced inequality ratio	P₊ =	0.214 135
Indegree balanced inequality ratio	P₋ =	0.304 024
Relative edge distribution entropy	H_er =	0.912 734
Power law exponent	γ =	1.744 92
Tail power law exponent	γ_t =	2.231 00
Degree assortativity	ρ =	−0.057 003 6
Degree assortativity p-value	p_ρ =	0.000 00
In/outdegree correlation	ρ^± =	+0.397 732
Clustering coefficient	c =	0.014 448 5
Directed clustering coefficient	c^± =	0.187 912
Spectral norm	α =	349.929
Operator 2-norm	ν =	336.756
Cyclic eigenvalue	π =	89.433 7
Algebraic connectivity	a =	0.000 162 332
Reciprocity	y =	0.343 240
Non-bipartivity	b_A =	0.043 241 4
Normalized non-bipartivity	b_N =	0.000 182 942
Algebraic non-bipartivity	χ =	0.000 404 033
Spectral bipartite frustration	b_K =	1.183 14 × 10⁻⁵
Controllability	C =	914,142
Relative controllability	C_r =	0.570 701

[1]	Jérôme Kunegis. KONECT – The Koblenz Network Collection. In Proc. Int. Conf. on World Wide Web Companion, pages 1343–1350, 2013. [ http ]
[2]	Peter Bailey, Nick Craswell, and David Hawking. Engineering a multi-purpose test collection for Web retrieval experiments. Inf. Process. and Manag., 39(6):853–871, 2003.