Multi-domain and Explainable Prediction of Changes in Web Vocabularies

This page contains additional data and analysis supporting the paper:

• Characteristics of input datasets
• Selected features per dataset
• Finding relationships between features and classifier performance
• Finding relationships between features and best classifier

The following table describes the 139 datasets evalauted in the paper. Each dataset consists of exactly one version chain. One version chain consists of a varying number of versions (at least 3). Datasets 1 and 2 are the CEDAR and DBpedia datasets. Datasets 3-136 are the LOV datasets (i.e. retrieved via the Linked Open Vocabularies API). Datasets 137-139 are the SPARQL datasets (i.e. version chains reconstructed after querying the 637 public SPARQL endpoints in datahub.io). Each dataset is described by the following characteristics/features:

• dataset: name of the version chain
• totalSize: total size of the version chain, adding up the total number of triples in all versions
• nSnapshots: number of versions/snapshots
• avgGap: average time gap between the versions (in number of days) considering their release date
• avgSize: average size of the version chain, in number of triples
• nInserts: total number of inserted triples between all version pairs
• nDeletes: total number of deleted triples between all version pairs
• nComm: total number of common/stable triples between all version pairs
• ratioInserts: ratio of inter-version statement insertions versus total number of inter-version operations
• ratioDeletes: ratio of inter-version statement deletions versus total number of inter-version operations
• ratioComm: ratio of inter-version common statements versus total number of inter-version operations
• isTree: whether the schema forms a tree or a graph
• maxTreeDepth: maximum tree depth considering all versions
• avgTreeDepth: average tree depth
• totalInstances: total number of instances in the version chain
• ratioInstances: ratio of instance triples versus total number of triples
• ratioInstancesVSIS: ratio of instance triples versus instance and structural triples
• totalStructural: total number of structural properties in the version chain
• ratioStructural: ratio of structural triples versus total number of triples
• ratioStructuralVSIS: ratio of structural triples versus instance and structural triples

	totalSize	nSnapshots	avgGap	avgSize	nInserts	nDeletes	nComm	ratioInserts	ratioDeletes	ratioComm	maxTreeDepth	avgTreeDepth	totalInstances	ratioInstances	ratioInstancesVSIS	totalStructural	ratioStructural	ratioStructuralVSIS
1	19639805	8	4223.57	2454975.62	17420246	19043105	2052	0.48	0.52	0.00	3	3.00	2173259	0.11	1.00	3585	0.00	0.00
2	50084545	5	278.75	10016909.00	15269985	4862044	29313340	0.31	0.10	0.59	2	2.00	50049049	1.00	1.00	2049	0.00	0.00
3	736	4	181.33	184.00	233	367	222	0.28	0.45	0.27	1	1.00	121	0.16	0.91	12	0.02	0.09
4	776	3	67.00	258.67	384	250	221	0.45	0.29	0.26	1	1.00	110	0.14	0.75	37	0.05	0.25
5	3281	3	99.50	1093.67	1389	1861	631	0.36	0.48	0.16	1	1.00	1093	0.33	0.64	603	0.18	0.36
6	1395	3	1607.00	465.00	444	660	344	0.31	0.46	0.24	3	3.00	296	0.21	0.80	72	0.05	0.20
7	1853	3	61.50	617.67	397	470	817	0.24	0.28	0.49	1	1.67	230	0.12	0.73	86	0.05	0.27
8	1108	5	203.00	221.60	344	274	605	0.28	0.22	0.49	1	1.00	190	0.17	0.67	94	0.08	0.33
9	518	4	256.67	129.50	82	39	328	0.18	0.09	0.73	1	1.00	94	0.18	0.84	18	0.03	0.16
10	351	3	385.50	117.00	53	30	195	0.19	0.11	0.70	2	1.67	66	0.19	0.61	42	0.12	0.39
11	543	4	256.67	135.75	82	37	345	0.18	0.08	0.74	2	2.00	93	0.17	0.79	24	0.04	0.21
12	330	3	243.50	110.00	37	22	191	0.15	0.09	0.76	2	2.00	53	0.16	0.78	15	0.05	0.22
13	6896	4	104.33	1724.00	1369	1359	3903	0.21	0.20	0.59	5	5.00	935	0.14	0.67	453	0.07	0.33
14	4266	3	259.50	1422.00	1398	1742	1310	0.31	0.39	0.29	3	2.33	844	0.20	0.76	264	0.06	0.24
15	2371	7	502.00	338.71	1432	513	895	0.50	0.18	0.32	3	1.57	339	0.14	0.76	106	0.04	0.24
16	5637	1	298.00	5637.00							7	8.00	1200	0.21	0.72	470	0.08	0.28
17	387	3	368.50	129.00	117	136	130	0.31	0.36	0.34	2	2.00	119	0.31	0.69	54	0.14	0.31
18	1027	4	449.00	256.75	338	457	378	0.29	0.39	0.32	2	2.00	293	0.29	0.83	58	0.06	0.17
19	744	3	789.00	248.00	69	80	428	0.12	0.14	0.74	1	1.00	158	0.21	0.96	6	0.01	0.04
20	139	3	1177.50	46.33	27	30	67	0.22	0.24	0.54	1	1.33	12	0.09	0.36	21	0.15	0.64
21	5455	12	137.82	454.58	709	1196	3768	0.12	0.21	0.66	3	0.25	1237	0.23	0.93	94	0.02	0.07
22	547	3	486.50	182.33	329	438	2	0.43	0.57	0.00	1	1.00	143	0.26	0.66	73	0.13	0.34
23	499	6	329.40	83.17	74	69	356	0.15	0.14	0.71	1	1.00	118	0.24	0.84	23	0.05	0.16
24	3134	4	126.00	783.50	119	142	2241	0.05	0.06	0.90	4	2.75	1243	0.40	0.70	528	0.17	0.30
25	672	5	100.60	134.40	174	104	387	0.26	0.16	0.58	2	1.80	174	0.26	0.81	41	0.06	0.19
26	303	5	226.25	60.60	47	43	201	0.16	0.15	0.69	1	1.00	79	0.26	0.83	16	0.05	0.17
27	10017	6	615.20	1669.50	6986	7500	1297	0.44	0.48	0.08	4	3.33	2395	0.24	0.55	1993	0.20	0.45
28	1796	9	352.38	199.56	398	138	1330	0.21	0.07	0.71	2	1.11	510	0.28	0.94	34	0.02	0.06
29	59	3	806.00	19.67	17	16	25	0.29	0.28	0.43	0	0.00	6	0.10	1.00	0	0.00	0.00
30	2092	6	153.20	348.67	534	221	1450	0.24	0.10	0.66	1	1.00	258	0.12	0.91	24	0.01	0.09
31	418	3	806.00	139.33	97	98	183	0.26	0.26	0.48	0	0.00	45	0.11	1.00	0	0.00	0.00
32	2263	9	130.67	251.44	891	579	1349	0.32	0.21	0.48	3	1.44	809	0.36	0.90	93	0.04	0.10
33	2592	3	806.00	864.00	460	451	1276	0.21	0.21	0.58	1	1.00	296	0.11	0.92	27	0.01	0.08
34	343	3	806.00	114.33	67	68	163	0.22	0.23	0.55	1	1.00	36	0.10	0.86	6	0.02	0.14
35	1697	3	173.00	565.67	380	513	668	0.24	0.33	0.43	1	1.00	381	0.22	0.95	21	0.01	0.05
36	87839	7	42.43	12548.43	41513	45645	30183	0.35	0.39	0.26	6	4.29	18942	0.22	0.50	18972	0.22	0.50
37	1414	3	335.50	471.33	353	481	505	0.26	0.36	0.38	2	2.00	273	0.19	0.58	195	0.14	0.42
38	707	6	254.00	117.83	396	312	250	0.41	0.33	0.26	1	1.00	99	0.14	0.94	6	0.01	0.06
39	11736	6	370.80	1956.00	3601	3710	6168	0.27	0.28	0.46	4	3.33	2139	0.18	0.64	1180	0.10	0.36
40	2051	7	611.33	293.00	1902	1812	82	0.50	0.48	0.02	1	0.86	460	0.22	0.83	96	0.05	0.17
41	17470	7	118.00	2495.71	6144	5584	9006	0.30	0.27	0.43	4	3.29	3444	0.20	0.81	792	0.05	0.19
42	1710	4	470.00	427.50	684	614	632	0.35	0.32	0.33	3	3.00	306	0.18	0.86	50	0.03	0.14
43	658	3	126.00	219.33	73	30	396	0.15	0.06	0.79	1	1.00	103	0.16	0.79	27	0.04	0.21
44	1359	4	287.33	339.75	121	116	909	0.11	0.10	0.79	1	1.00	298	0.22	0.96	12	0.01	0.04
45	800	3	335.50	266.67	178	274	296	0.24	0.37	0.40	1	1.00	161	0.20	0.62	98	0.12	0.38
46	29393	15	91.79	1959.53	8383	8097	19313	0.23	0.23	0.54	5	3.67	7359	0.25	0.62	4601	0.16	0.38
47	5624	10	356.22	562.40	276	111	4889	0.05	0.02	0.93	2	2.00	1440	0.26	0.91	137	0.02	0.09
48	1393	6	61.00	232.17	251	24	1023	0.19	0.02	0.79	2	1.83	500	0.36	0.95	24	0.02	0.05
49	870	3	82.00	290.00	195	267	341	0.24	0.33	0.42	2	2.00	166	0.19	0.82	37	0.04	0.18
50	1561	4	55.00	390.25	444	576	634	0.27	0.35	0.38	3	3.25	301	0.19	0.69	135	0.09	0.31
51	6670	3	691.50	2223.33	641	641	3863	0.12	0.12	0.75	2	2.33	2573	0.39	0.92	216	0.03	0.08
52	616	3	347.00	205.33	166	221	210	0.28	0.37	0.35	3	2.00	98	0.16	0.60	66	0.11	0.40
53	45667	12	274.45	3805.58	7632	3445	35386	0.16	0.07	0.76	2	1.25	8716	0.19	0.99	120	0.00	0.01
54	15959	4	231.67	3989.75	5526	6665	5588	0.31	0.37	0.31	3	2.75	2008	0.13	0.90	225	0.01	0.10
55	8803	3	365.50	2934.33	655	646	5224	0.10	0.10	0.80	4	5.00	1767	0.20	0.54	1485	0.17	0.46
56	253	3	475.50	84.33	36	56	123	0.17	0.26	0.57	1	0.33	44	0.17	0.83	9	0.04	0.17
57	164	3	661.00	54.67	92	39	54	0.50	0.21	0.29	1	0.67	35	0.21	0.95	2	0.01	0.05
58	934	3	120.00	311.33	78	54	557	0.11	0.08	0.81	3	2.33	144	0.15	0.71	59	0.06	0.29
59	1997	3	82.50	665.67	453	620	768	0.25	0.34	0.42	2	2.00	406	0.20	0.84	79	0.04	0.16
60	3552	5	39.00	710.40	384	526	2318	0.12	0.16	0.72	5	2.80	648	0.18	0.70	273	0.08	0.30
61	376	3	376.50	125.33	68	45	200	0.22	0.14	0.64	1	1.00	89	0.24	0.93	7	0.02	0.07
62	367	3	337.50	122.33	190	82	106	0.50	0.22	0.28	3	2.00	58	0.16	0.79	15	0.04	0.21
63	3397	4	302.00	849.25	1730	1299	1139	0.42	0.31	0.27	2	1.50	663	0.20	0.76	212	0.06	0.24
64	17391	6	376.80	2898.50	10527	5808	6527	0.46	0.25	0.29	4	1.83	4913	0.28	0.85	843	0.05	0.15
65	74610	4	402.33	18652.50	29449	17608	36442	0.35	0.21	0.44	1	1.25	7242	0.10	0.59	4944	0.07	0.41
66	691	3	375.00	230.33	348	321	168	0.42	0.38	0.20	0	0.00	143	0.21	0.63	85	0.12	0.37
67	6206	13	205.38	477.38	1084	338	5021	0.17	0.05	0.78	1	0.85	563	0.09	0.95	27	0.00	0.05
68	1818	4	112.50	454.50	951	1104	411	0.39	0.45	0.17	3	2.00	473	0.26	0.78	134	0.07	0.22
69	406	3	485.00	135.33	61	39	226	0.19	0.12	0.69	1	1.00	79	0.19	0.73	29	0.07	0.27
70	1012	16	57.60	63.25	259	143	746	0.23	0.12	0.65	1	0.31	146	0.14	0.92	13	0.01	0.08
71	782	4	272.33	195.50	41	16	556	0.07	0.03	0.91	1	1.00	151	0.19	0.97	4	0.01	0.03
72	570	4	481.33	142.50	273	190	213	0.40	0.28	0.32	2	1.50	163	0.29	0.95	8	0.01	0.05
73	11308	4	254.67	2827.00	5859	5434	3399	0.40	0.37	0.23	2	2.00	2418	0.21	0.61	1551	0.14	0.39
74	2211	3	168.00	737.00	406	403	1071	0.22	0.21	0.57	2	2.33	577	0.26	0.73	209	0.09	0.27
75	42866	5	239.25	8573.20	8853	8591	26074	0.20	0.20	0.60	0	0.00	3718	0.09	1.00	0	0.00	0.00
76	21094	7	153.33	3013.43	4466	3962	13663	0.20	0.18	0.62	0	0.00	7327	0.35	1.00	0	0.00	0.00
77	309	4	589.00	77.25	140	87	120	0.40	0.25	0.35	1	1.00	72	0.23	0.91	7	0.02	0.09
78	3008	3	51.00	1002.67	1134	720	1149	0.38	0.24	0.38	5	4.67	509	0.17	0.65	277	0.09	0.35
79	2802	3	1565.00	934.00	1363	765	951	0.44	0.25	0.31	3	3.00	619	0.22	0.61	388	0.14	0.39
80	4062	10	52.44	406.20	359	265	3337	0.09	0.07	0.84	2	2.20	799	0.20	0.83	159	0.04	0.17
81	320	2	192.00	160.00	19	9	147	0.11	0.05	0.84	1	1.00	43	0.13	0.91	4	0.01	0.09
82	1318	3	174.50	439.33	1011	757	2	0.57	0.43	0.00	2	2.00	137	0.10	0.64	76	0.06	0.36
83	5244	4	230.00	1311.00	2754	2365	1598	0.41	0.35	0.24	1	1.00	873	0.17	0.84	164	0.03	0.16
84	162	3	6.50	54.00	97	57	32	0.52	0.31	0.17	0	0.00	41	0.25	1.00	0	0.00	0.00
85	394	3	6.00	131.33	84	135	159	0.22	0.36	0.42	1	1.00	105	0.27	0.75	35	0.09	0.25
86	458	4	184.33	114.50	22	20	326	0.06	0.05	0.89	1	1.00	124	0.27	0.97	4	0.01	0.03
87	529	3	81.50	176.33	110	136	223	0.23	0.29	0.48	3	2.00	82	0.16	0.59	56	0.11	0.41
88	858	4	339.00	214.50	122	95	533	0.16	0.13	0.71	2	1.50	234	0.27	0.84	46	0.05	0.16
89	5202	10	156.11	520.20	891	453	4009	0.17	0.08	0.75	2	2.40	932	0.18	0.92	81	0.02	0.08
90	1403	3	339.00	467.67	42	14	910	0.04	0.01	0.94	0	0.00	237	0.17	1.00	0	0.00	0.00
91	3651	3	497.00	1217.00	2197	2607	199	0.44	0.52	0.04	3	3.33	992	0.27	0.92	85	0.02	0.08
92	1820	8	293.00	227.50	484	226	1255	0.25	0.12	0.64	0	0.00	328	0.18	1.00	0	0.00	0.00
93	338	3	107.00	112.67	54	68	164	0.19	0.24	0.57	1	1.00	57	0.17	0.90	6	0.02	0.10
94	1405	3	229.50	468.33	399	148	627	0.34	0.13	0.53	2	2.00	220	0.16	0.76	68	0.05	0.24
95	521	5	95.75	104.20	119	134	289	0.22	0.25	0.53	1	1.00	122	0.23	0.88	16	0.03	0.12
96	3078	11	112.50	279.82	947	781	2013	0.25	0.21	0.54	2	1.64	1143	0.37	0.90	127	0.04	0.10
97	7523	5	245.75	1504.60	2720	1983	3881	0.32	0.23	0.45	2	2.60	845	0.11	0.78	235	0.03	0.22
98	2732	8	280.57	341.50	936	887	1554	0.28	0.26	0.46	3	2.50	703	0.26	0.79	188	0.07	0.21
99	2050	8	170.86	256.25	663	487	1259	0.28	0.20	0.52	2	1.25	787	0.38	0.90	90	0.04	0.10
100	628	3	619.00	209.33	219	210	205	0.35	0.33	0.32	1	0.33	205	0.33	0.93	16	0.03	0.07
101	1033	4	447.00	258.25	96	64	706	0.11	0.07	0.82	3	2.25	280	0.27	0.84	53	0.05	0.16
102	8903	4	235.33	2225.75	188	186	6492	0.03	0.03	0.95	0	0.00	1004	0.11	1.00	0	0.00	0.00
103	520	3	290.00	173.33	48	16	314	0.13	0.04	0.83	0	0.00	137	0.26	1.00	0	0.00	0.00
104	1098	3	277.50	366.00	309	513	289	0.28	0.46	0.26	1	1.33	261	0.24	0.72	101	0.09	0.28
105	1058	3	196.50	352.67	654	325	263	0.53	0.26	0.21	2	1.33	292	0.28	0.77	88	0.08	0.23
106	1266	3	382.00	422.00	334	460	429	0.27	0.38	0.35	2	2.33	216	0.17	0.59	151	0.12	0.41
107	4320	2	149.00	2160.00	1427	2893	1	0.33	0.67	0.00	6	3.50	1185	0.27	0.85	212	0.05	0.15
108	129601	20	50.32	6480.05	7417	1885	119431	0.06	0.01	0.93	3	2.50	27090	0.21	0.71	11088	0.09	0.29
109	420	4	790.33	105.00	111	50	255	0.27	0.12	0.61	2	1.75	116	0.28	0.80	29	0.07	0.20
110	17608	3	47.50	5869.33	7950	11915	1147	0.38	0.57	0.05	2	1.33	3948	0.22	0.98	82	0.00	0.02
111	108767	7	67.00	15538.14	27445	32918	61517	0.23	0.27	0.50	9	5.29	17655	0.16	0.57	13091	0.12	0.43
112	129	3	1020.00	43.00	29	25	63	0.25	0.21	0.54	0	0.00	46	0.36	0.90	5	0.04	0.10
113	379	3	92.50	126.33	10	9	245	0.04	0.03	0.93	1	1.00	87	0.23	0.96	4	0.01	0.04
114	932	3	641.00	310.67	169	93	490	0.22	0.12	0.65	3	2.67	191	0.20	0.69	85	0.09	0.31
115	1779	4	435.33	444.75	536	352	932	0.29	0.19	0.51	3	2.25	380	0.21	0.83	77	0.04	0.17
116	2216	5	156.00	443.20	337	172	1522	0.17	0.08	0.75	2	2.20	569	0.26	0.74	198	0.09	0.26
117	4644	3	57.50	1548.00	1240	1476	1668	0.28	0.34	0.38	5	2.67	1052	0.23	0.51	1006	0.22	0.49
118	2435	3	36.00	811.67	355	271	1302	0.18	0.14	0.68	4	4.00	547	0.22	0.66	288	0.12	0.34
119	296	3	1653.50	98.67	26	26	176	0.11	0.11	0.77	1	1.00	53	0.18	0.79	14	0.05	0.21
120	1806	5	213.25	361.20	84	27	1387	0.06	0.02	0.93	4	2.40	493	0.27	0.71	204	0.11	0.29
121	1215	6	225.40	202.50	180	106	880	0.15	0.09	0.75	0	0.00	221	0.18	1.00	0	0.00	0.00
122	820	5	174.50	164.00	362	340	322	0.35	0.33	0.31	1	1.00	128	0.16	0.64	73	0.09	0.36
123	166	3	1007.00	55.33	38	37	78	0.25	0.24	0.51	0	0.00	51	0.31	0.84	10	0.06	0.16
124	210	3	1025.50	70.00	56	36	102	0.29	0.19	0.53	1	1.00	69	0.33	0.84	13	0.06	0.16
125	1273	5	289.25	254.60	140	124	862	0.12	0.11	0.77	0	0.00	271	0.21	1.00	0	0.00	0.00
126	2309	4	249.67	577.25	1305	1274	484	0.43	0.42	0.16	4	3.75	464	0.20	0.60	312	0.14	0.40
127	7478	3	82.50	2492.67	1308	1712	3485	0.20	0.26	0.54	2	1.00	1636	0.22	0.92	142	0.02	0.08
128	12771	8	51.43	1596.38	259	165	10954	0.02	0.01	0.96	2	2.00	3467	0.27	0.95	179	0.01	0.05
129	14625	8	331.29	1828.12	12553	10984	1945	0.49	0.43	0.08	1	1.00	943	0.06	0.80	237	0.02	0.20
130	5650	2	415.00	2825.00	1443	1443	1383	0.34	0.34	0.32	2	2.00	1406	0.25	0.79	382	0.07	0.21
131	2266	5	284.75	453.20	1134	908	797	0.40	0.32	0.28	3	1.20	476	0.21	0.74	171	0.08	0.26
132	4505	6	967.20	750.83	3926	3194	445	0.52	0.42	0.06	2	1.17	1093	0.24	0.80	281	0.06	0.20
133	11196	4	1434.67	2799.00	7467	8779	580	0.44	0.52	0.03	3	2.75	2559	0.23	0.75	839	0.07	0.25
134	56130	10	186.11	5613.00	26124	24157	25200	0.35	0.32	0.33	6	4.10	12120	0.22	0.66	6322	0.11	0.34
135	1662	7	233.50	237.43	493	300	1046	0.27	0.16	0.57	1	1.00	269	0.16	0.96	12	0.01	0.04
136	859	3	406.00	286.33	298	296	280	0.34	0.34	0.32	1	1.00	238	0.28	0.99	3	0.00	0.01
137	22404	3		7468.00	22347	22345	2	0.50	0.50	0.00	2	1.00	10789	0.48	1.00	3	0.00	0.00
138	16505	3		5501.67	196	4147	9464	0.01	0.30	0.69	2	1.67	2259	0.14	0.98	37	0.00	0.02
139	1431	5		286.20	196	951	380	0.13	0.62	0.25	1	0.60	196	0.14	0.98	5	0.00	0.02

The following table shows the top 10 ranked features by the RELIEF feature selection algorithm for LOV and SPARQL datasets. A 0 means that no further significant features were selected. Other integer values correspond to the features discussed in the paper with the following equivalences:

• 1: dirChildren
• 2: dirChildren1
• 3: dirChildren2
• 4: dirChildren3
• 5: numParents
• 6: numSiblings
• 7: dirArticles
• 8: dirArticlesChildren0
• 9: dirArticlesChildren1
• 10: dirArticlesChildren2
• 11: dirArticlesChildren3
• 12: ratio
• 13: ratio0
• 14: ratio1
• 15: ratio2
• 16: ratio3

	dataset	1stFeat	2ndFeat	3rdFeat	4thFeat	5thFeat	6thFeat	7thFeat	8thFeat	9thFeat	10thFeat
1	lov/adms-4-1-1-allDrift-T	2	3	4	1	6	5	7	15	16	8
2	lov/aiiso-3-1-1-allDrift-T	5	6	0	0	0	0	0	0	0	0
3	lov/bag-3-1-1-allDrift-T	5	6	0	0	0	0	0	0	0	0
4	lov/basic-5-1-1-allDrift-T	6	4	3	2	1	11	10	12	16	15
5	lov/bbc-4-1-1-allDrift-T	1	3	2	4	5	6	12	13	8	7
6	lov/bbccms-3-1-1-allDrift-T	5	6	0	0	0	0	0	0	0	0
7	lov/bbccore-4-1-1-allDrift-T	1	4	2	3	5	6	10	12	13	11
8	lov/bibo-3-1-1-allDrift-T	6	5	0	0	0	0	0	0	0	0
9	lov/bio-7-1-1-allDrift-T	6	5	5	6	6	5	8	10	9	14
10	lov/biro-3-1-1-allDrift-T	3	4	2	1	0	0	0	0	0	0
11	lov/c4o-4-1-1-allDrift-T	2	3	4	1	6	5	7	15	16	8
12	lov/cnt-3-1-1-allDrift-T	16	15	7	10	9	8	11	14	13	12
13	lov/co-6-1-1-allDrift-T	5	6	4	1	3	2	15	14	16	11
14	lov/cogs-4-1-1-allDrift-T	6	6	5	5	0	0	0	0	0	0
15	lov/cold-5-1-1-allDrift-T	8	16	9	10	15	11	14	13	7	12
16	lov/comm-5-1-1-allDrift-T	1	4	2	3	5	6	4	3	2	1
17	lov/d2rq-9-1-1-allDrift-T	6	6	6	5	5	5	1	3	2	4
18	lov/dcat-6-1-1-allDrift-T	1	4	2	3	5	6	15	11	12	13
19	lov/dcite-9-1-1-allDrift-T	11	9	10	12	13	8	7	16	15	14
20	lov/earl-7-1-1-allDrift-T	15	16	14	11	10	9	8	7	13	12
21	lov/ebucore-7-1-1-allDrift-T	16	8	7	10	9	11	15	14	13	12
22	lov/edm-4-1-1-allDrift-T	2	3	4	1	15	7	8	16	10	13
23	lov/emp-4-1-1-allDrift-T	1	1	3	2	4	3	2	4	5	5
24	lov/fabio-15-1-1-allDrift-T	16	14	12	13	15	7	16	7	11	10
25	lov/foaf-10-1-1-allDrift-T	16	15	14	11	12	10	16	7	8	9
26	lov/food-6-1-1-allDrift-T	1	1	3	4	2	4	3	2	5	6
27	lov/geofla-3-1-1-allDrift-T	3	4	2	1	0	0	0	0	0	0
28	lov/geom-4-1-1-allDrift-T	5	6	5	6	13	16	12	14	10	11
29	lov/gn-12-1-1-allDrift-T	5	7	12	8	13	6	0	0	0	0
30	lov/gndo-4-1-1-allDrift-T	16	8	15	10	9	7	11	14	13	12
31	lov/hr-3-1-1-allDrift-T	16	15	7	10	9	8	11	13	14	12
32	lov/itsmo-5-1-1-allDrift-T	13	8	7	12	14	15	16	12	13	7
33	lov/ldp-3-1-1-allDrift-T	6	5	0	0	0	0	0	0	0	0
34	lov/lemon-4-1-1-allDrift-T	6	6	5	5	0	0	0	0	0	0
35	lov/lexinfo-6-1-1-allDrift-T	6	16	8	10	9	11	7	12	13	14
36	lov/lgdo-4-1-1-allDrift-T	16	7	15	10	8	9	11	12	13	14
37	lov/lingvo-13-1-1-allDrift-T	16	8	15	10	9	7	11	14	12	13
38	lov/lv-16-1-1-allDrift-T	11	14	13	15	9	16	12	10	8	7
39	lov/marl-4-1-1-allDrift-T	16	8	15	10	9	7	11	13	14	12
40	lov/md-4-1-1-allDrift-T	16	8	15	10	9	7	11	12	14	13
41	lov/mrel-7-1-1-allDrift-T	16	7	15	14	13	16	1	2	5	4
42	lov/msm-4-1-1-allDrift-T	16	13	14	11	12	10	15	7	8	9
43	lov/mtlo-3-1-1-allDrift-T	5	6	3	2	0	0	0	0	0	0
44	lov/music-3-1-1-allDrift-T	5	6	4	3	1	0	0	0	0	0
45	lov/nif-10-1-1-allDrift-T	5	5	5	5	3	4	13	12	9	10
46	lov/ntag-3-1-1-allDrift-T	6	5	0	0	0	0	0	0	0	0
47	lov/ocd-4-1-1-allDrift-T	1	3	4	2	12	13	10	11	8	9
48	lov/opmw-4-1-1-allDrift-T	5	10	8	7	15	9	16	11	14	13
49	lov/org-10-1-1-allDrift-T	8	7	12	13	3	4	14	16	9	10
50	lov/pattern-3-1-1-allDrift-T	6	1	2	5	3	4	0	0	0	0
51	lov/pav-8-1-1-allDrift-T	16	8	15	10	9	7	11	13	14	12
52	lov/po-3-1-1-allDrift-T	5	6	0	0	0	0	0	0	0	0
53	lov/poste-5-1-1-allDrift-T	1	3	4	2	5	6	7	15	16	8
54	lov/pro-11-1-1-allDrift-T	1	4	3	2	5	6	13	8	15	14
55	lov/prov-5-1-1-allDrift-T	4	3	2	1	13	12	8	16	10	11
56	lov/prv-8-1-1-allDrift-T	1	2	1	2	0	0	0	0	0	0
57	lov/pso-8-1-1-allDrift-T	1	3	2	4	6	5	11	13	14	15
58	lov/qb-4-1-1-allDrift-T	8	12	7	13	16	15	14	1	9	10
59	lov/ruto-3-1-1-allDrift-T	16	15	7	10	9	8	11	13	14	12
60	lov/sam-3-1-1-allDrift-T	1	2	3	4	6	0	0	0	0	0
61	lov/semio-4-1-1-allDrift-T	5	5	6	1	4	3	2	6	15	16
62	lov/sio-7-1-1-allDrift-T	16	10	9	12	11	13	7	16	15	14
63	lov/sport-4-1-1-allDrift-T	5	5	6	6	1	2	3	4	15	16
64	lov/spt-5-1-1-allDrift-T	16	13	14	11	12	10	16	7	15	9
65	lov/taxon-5-1-1-allDrift-T	5	5	9	11	12	13	14	10	8	16
66	lov/teach-6-1-1-allDrift-T	1	4	6	5	3	2	14	13	16	11
67	lov/thors-5-1-1-allDrift-T	6	5	4	1	3	2	12	13	14	15
68	lov/tisc-5-1-1-allDrift-T	3	1	4	5	6	2	1	6	3	5
69	lov/tm-4-1-1-allDrift-T	6	5	4	6	0	0	0	0	0	0
70	lov/txn-8-1-1-allDrift-T	5	5	5	5	9	8	10	16	11	12
71	lov/umbel-8-1-1-allDrift-T	6	6	6	6	5	5	5	5	12	11
72	lov/vivo-10-1-1-allDrift-T	1	1	6	1	6	1	1	6	6	2
73	lov/voaf-7-1-1-allDrift-T	5	6	4	5	2	3	6	1	1	4
74	sparql/fao-3-1-1-allDrift-T	16	15	7	10	9	8	11	12	13	14
75	sparql/lingvoj-5-1-1-allDrift-T	6	5	10	15	14	9	13	12	16	8

Through regression, we analyse what dataset characteristics are good predictors of the performance of the best selected classifier in our approach, using the area under the ROC curve as a response variable. We find that, under the null hypothesis of normality and non-dependence, the predictors nSnapshots, avgTreeDepth, ratioStructural, ratioInserts and ratioComm are good explanatory variables (i.e. have an influence) with respect to the performance of change detection in version chains. The first model, which includes ratioInserts discarding ratioDeletes and ratioComm due to multi-colinearity, shows the best model fit with respect to the data. The figure below depicts this.

	Dependent variable:
	roc
	(1)	(2)	(3)
log(nSnapshots)	0.365***	0.350***	0.370***
	(0.080)	(0.085)	(0.083)
log(avgGap)	-0.001	0.013	0.006
	(0.031)	(0.032)	(0.032)
log(totalSize)	-0.029	-0.023	-0.029
	(0.021)	(0.023)	(0.022)
avgTreeDepth	0.114***	0.113***	0.115***
	(0.038)	(0.039)	(0.038)
ratioInstances	0.465	0.426	0.477
	(0.327)	(0.343)	(0.336)
ratioStructural	-1.858**	-1.711**	-1.895**
	(0.755)	(0.804)	(0.782)
ratioInserts	0.748***
	(0.249)
ratioDeletes		0.173
		(0.253)
ratioComm			-0.265*
			(0.136)
Constant	-0.125	-0.064	0.154
	(0.240)	(0.249)	(0.261)
Observations	131	131	131
R2	0.269	0.218	0.239
Adjusted R2	0.228	0.174	0.196
Residual Std. Error (df = 123)	0.345	0.357	0.352
F Statistic (df = 7; 123)	6.471***	4.908***	5.517***
Note:	*p<0.1; **p<0.05; ***p<0.01

Through multinomial logistic regression, we analyse what dataset characteristics are good predictors of the classifier type selected as best in our approach. We find that avgGap is influential at selecting a tree classifier instead of a bayes one. We also find that totalSize is influential at selecting functions and rules based classifiers instead of bayes ones. The pictures below show simulations on how these predictors influcence the choice of the different classifier families, in one overall figure and two detailed ones (simulating the effect of avgGap and totalSize).

	Dependent variable:
	functions	rules	trees	functions	rules	trees	functions	rules	trees
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
log(nSnapshots)	-0.291	-0.257	1.975	-0.180	-0.239	1.745	-0.193	-0.212	1.838
	(0.656)	(0.765)	(1.503)	(0.680)	(0.790)	(1.512)	(0.667)	(0.777)	(1.497)
log(avgGap)	0.238	0.145	1.385*	0.266	0.173	1.269*	0.248	0.161	1.351*
	(0.242)	(0.271)	(0.734)	(0.240)	(0.269)	(0.703)	(0.240)	(0.270)	(0.729)
log(totalSize)	0.669***	0.539*	-0.052	0.636**	0.531*	-0.010	0.641***	0.524*	-0.025
	(0.249)	(0.278)	(0.563)	(0.251)	(0.282)	(0.555)	(0.249)	(0.279)	(0.557)
avgTreeDepth	-0.399	-0.334	0.534	-0.393	-0.336	0.564	-0.385	-0.323	0.553
	(0.302)	(0.330)	(0.719)	(0.304)	(0.334)	(0.728)	(0.303)	(0.332)	(0.728)
ratioInstances	1.378	2.463	3.090	1.071	2.246	3.394	1.269	2.330	3.221
	(3.485)	(4.021)	(6.654)	(3.455)	(3.981)	(6.629)	(3.476)	(4.005)	(6.649)
ratioStructural	-9.054	1.357	-9.539	-9.039	1.674	-10.799	-9.594	1.116	-10.030
	(6.040)	(6.135)	(13.505)	(6.142)	(6.353)	(13.945)	(6.136)	(6.267)	(13.827)
ratioInserts	3.006	2.376	-3.540
	(1.906)	(2.210)	(4.401)
ratioDeletes				1.918	0.929	-2.341
				(1.907)	(2.154)	(4.058)
ratioComm							-1.440	-0.945	1.615
							(1.028)	(1.170)	(2.219)
Constant	-5.610**	-5.580**	-12.702**	-5.288**	-5.259**	-12.402**	-4.059*	-4.494*	-14.266**
	(2.248)	(2.511)	(5.954)	(2.210)	(2.494)	(5.759)	(2.265)	(2.585)	(6.511)
Akaike Inf. Crit.	313.543	313.543	313.543	316.179	316.179	316.179	314.605	314.605	314.605
Note:	*p<0.1; **p<0.05; ***p<0.01

---