Project

General

Profile

1
{
2
	"wf" : {
3
		"threshold" : "0.99",
4
		"dedupRun" : "001",
5
		"entityType" : "result",
6
		"orderField" : "title",
7
		"queueMaxSize" : "4000",
8
		"groupMaxSize" : "40",
9
		"slidingWindowSize" : "200",
10
		"rootBuilder" : [ "result", "personResult_authorship_hasAuthor", "resultProject_outcome_isProducedBy", "resultResult_publicationDataset_isRelatedTo", "resultResult_similarity_isAmongTopNSimilarDocuments", "resultResult_similarity_hasAmongTopNSimilarDocuments" ],
11
		"includeChildren" : "true",
12
		"maxChildren" : "40"
13
	},
14
	"pace" : {
15
		"clustering" : [
16
			{ "name" : "ngrampairs", "fields" : [ "title" ], "params" : { "max" : "1", "ngramLen" : "3"} },
17
			{ "name" : "suffixprefix", "fields" : [ "title" ], "params" : { "max" : "1", "len" : "3" } },
18
			{ "name" : "lowercase", "fields" : [ "doi" ], "params" : { } }
19
		],
20
		"strictConditions" : [
21
			{ "name" : "pidMatch", "fields" : [ "pid" ] }
22
		],
23
		"conditions" : [
24
			{ "name" : "titleVersionMatch", "fields" : [ "title" ] },
25
			{ "name" : "sizeMatch", "fields" : [ "authors" ] }
26
		],
27
		"model" : [
28
			{ "name" : "doi", "algo" : "Null", "type" : "String", "weight" : "0.0", "ignoreMissing" : "true", "path" : "pid[qualifier#classid = {doi}]/value" },
29
			{ "name" : "pid", "algo" : "Null", "type" : "JSON", "weight" : "0.0", "ignoreMissing" : "true", "path" : "pid", "overrideMatch" : "true" },
30
			{ "name" : "title", "algo" : "LevensteinTitle", "type" : "String", "weight" : "1.0", "ignoreMissing" : "false", "path" : "result/metadata/title[qualifier#classid = {main title}]/value" },
31
			{ "name" : "authors", "algo" : "Null", "type" : "String", "weight" : "0.0", "ignoreMissing" : "true", "path" : "result/author/metadata/fullname/value" }
32
		],
33
		"blacklists" : {
34
			"title" : [
35
				"^THE ASSOCIATION AND THE GENERAL MEDICAL COUNCIL$",
36
				"^Problems with perinatal pathology\.?$",
37
				"(?i)^Cases? of Puerperal Convulsions$",
38
				"(?i)^Operative Gyna?ecology$",
39
				"(?i)^Mind the gap\!?\:?$",
40
				"^Chronic fatigue syndrome\.?$",
41
				"^Cartas? ao editor Letters? to the Editor$",
42
				"^Note from the Editor$",
43
				"^Anesthesia Abstract$",
44

    
45
				"^Annual report$",
46
				"(?i)^“?THE RADICAL PREVENTION OF VENEREAL DISEASE\.?”?$",
47
				"(?i)^Graph and Table of Infectious Diseases?$",
48
				"^Presentation$",
49
				"(?i)^Reviews and Information on Publications$",
50
				"(?i)^PUBLIC HEALTH SERVICES?$",
51
				"(?i)^COMBINED TEXT-?BOOK OF OBSTETRICS AND GYN(Æ|ae)COLOGY$",
52
				"(?i)^Adrese autora$",
53
				"(?i)^Systematic Part .*\. Catalogus Fossilium Austriae, Band 2: Echinoidea neogenica$",
54
				"(?i)^Acknowledgement to Referees$",
55
				"(?i)^Behçet's disease\.?$",
56
				"(?i)^Isolation and identification of restriction endonuclease.*$",
57
				"(?i)^CEREBROVASCULAR DISEASES?.?$",
58
				"(?i)^Screening for abdominal aortic aneurysms?\.?$",
59
				"^Event management$",
60
				"(?i)^Breakfast and Crohn's disease.*\.?$",
61
				"^Cálculo de concentraciones en disoluciones acuosas. Ejercicio interactivo\..*\.$",
62
				"(?i)^Genetic and functional analyses of SHANK2 mutations suggest a multiple hit model of Autism spectrum disorders?\.?$",
63
				"^Gushi hakubutsugaku$",
64

    
65
				"^Starobosanski nadpisi u Bosni i Hercegovini \(.*\)$",
66
				"^Intestinal spirocha?etosis$",
67
				"^Treatment of Rodent Ulcer$",
68
				"(?i)^\W*Cloud Computing\W*$",
69
				"^Compendio mathematico : en que se contienen todas las materias mas principales de las Ciencias que tratan de la cantidad$",
70
				"^Free Communications, Poster Presentations: Session [A-F]$",
71

    
72
				"^“The Historical Aspects? of Quackery\.?”$",
73
				"^A designated centre for people with disabilities operated by St John of God Community Services (Limited|Ltd), Louth$",
74
				"^P(er|re)-Mile Premiums for Auto Insurance\\.?$",
75
				"(?i)^Case Report$",
76
				"^Boletín Informativo$",
77
				"(?i)^Glioblastoma Multiforme$",
78
				"(?i)^Nuevos táxones animales descritos en la península Ibérica y Macaronesia desde 1994 \\(.*\\)$",
79
				"^Zaměstnanecké výhody$",
80
				"(?i)^The Economics of Terrorism and Counter-Terrorism: A Survey \\(Part .*\\)$",
81
				"(?i)^Carotid body tumours?\\.?$",
82
				"(?i)^\\[Españoles en Francia : La condición Emigrante.*\\]$",
83
				"^Avant-propos$",
84
				"(?i)^St\. Patrick's Cathedral, Dublin, County Dublin - Head(s)? and Capital(s)?$",
85
				"(?i)^St\. Patrick's Cathedral, Dublin, County Dublin - Bases?$",
86
				"(?i)^PUBLIC HEALTH VERSUS THE STATE$",
87
				"^Viñetas de Cortázar$",
88
				"(?i)^Search for heavy neutrinos and W(\\[|_|\\(|_\\{|-)?R(\\]|\\)|\\})? bosons with right-handed couplings in a left-right symmetric model in pp collisions at.*TeV(\.)?$",
89
				"(?i)^Measurement of the pseudorapidity and centrality dependence of the transverse energy density in Pb(-?)Pb collisions at.*tev(\.?)$",
90
				"(?i)^Search for resonances decaying into top-quark pairs using fully hadronic decays in pp collisions with ATLAS at.*TeV$",
91
				"(?i)^Search for neutral minimal supersymmetric standard model Higgs bosons decaying to tau pairs in pp collisions at.*tev$",
92

    
93
				"(?i)^Relatório de Estágio (de|em) Angiologia e Cirurgia Vascular$",
94
				"^Aus der AGMB$",
95

    
96
				"^Znanstveno-stručni prilozi$",
97
				"(?i)^Zhodnocení finanční situace podniku a návrhy na zlepšení$",
98
				"(?i)^Evaluation of the Financial Situation in the Firm and Proposals to its Improvement$",
99
				"(?i)^Hodnocení finanční situace podniku a návrhy na její zlepšení$",
100
				"^Finanční analýza podniku$",
101
				"^Financial analysis( of business)?$",
102
				"(?i)^Textbook of Gyn(a)?(Æ)?(e)?cology$",
103
				"^Jikken nihon shūshinsho$",
104
				"(?i)^CORONER('|s)(s|') INQUESTS$",
105
				"(?i)^(Μελέτη παραγόντων )?risk management( για ανάπτυξη και εφαρμογή ενός πληροφοριακού συστήματος| και ανάπτυξη συστήματος)?$",
106
				"(?i)^Consultants' contract(s)?$",
107
				"(?i)^Upute autorima$",
108
				"(?i)^Bijdrage tot de Kennis van den Godsdienst der Dajaks van Lan(d|f)ak en Tajan$",
109
				"^Joshi shin kokubun$",
110
				"^Kōtō shōgaku dokuhon nōson'yō$",
111
				"^Jinjō shōgaku shōka$",
112
				"^Shōgaku shūjichō$",
113
				"^Nihon joshi dokuhon$",
114
				"^Joshi shin dokuhon$",
115
				"^Chūtō kanbun dokuhon$",
116
				"^Wabun dokuhon$",
117
				"(?i)^(Analysis of economy selected village or town|Rozbor hospodaření vybrané obce či města)$",
118
				"(?i)^cardiac rehabilitation$",
119
				"(?i)^Analytical summary$",
120
				"^Thesaurus resolutionum Sacrae Congregationis Concilii$",
121
				"(?i)^Sumario analítico(\\s{1})?(Analitic summary)?$",
122
				"^Prikazi i osvrti$",
123
				"^Rodinný dům s provozovnou$",
124
				"^Family house with an establishment$",
125
				"^Shinsei chūtō shin kokugun$",
126
				"^Pulmonary alveolar proteinosis(\\.?)$",
127
				"^Shinshū kanbun$",
128
				"^Viñeta(s?) de Rodríguez$",
129
				"(?i)^RUBRIKA UREDNIKA$",
130
				"^A Matching Model of the Academic Publication Market$",
131
				"^Yōgaku kōyō$",
132

    
133
				"^Internetový marketing$",
134
				"^Internet marketing$",
135
				"^Chūtō kokugo dokuhon$",
136
				"^Kokugo dokuhon$",
137
				"^Antibiotic Cover for Dental Extraction(s?)$",
138
				"^Strategie podniku$",
139
				"^Strategy of an Enterprise$",
140
				"(?i)^respiratory disease(s?)(\.?)$",
141
				"^Award(s?) for Gallantry in Civil Defence$",
142
				"^Podniková kultura$",
143
				"^Corporate Culture$",
144
				"^Severe hyponatraemia in hospital inpatient(s?)(\.?)$",
145
				"^Pracovní motivace$",
146
				"^Work Motivation$",
147
				"^Kaitei kōtō jogaku dokuhon$",
148
				"^Konsolidovaná účetní závěrka$",
149
				"^Consolidated Financial Statements$",
150
				"(?i)^intracranial tumour(s?)$",
151
				"^Climate Change Mitigation Options and Directed Technical Change: A Decentralized Equilibrium Analysis$",
152
				"^\\[CERVECERIAS MAHOU(\\.|\\:) INTERIOR\\] \\[Material gráfico\\]$",
153
				"^Housing Market Dynamics(\\:|\\.) On the Contribution of Income Shocks and Credit Constraint(s?)$",
154
				"^\\[Funciones auxiliares de la música en Radio París,.*\\]$",
155
				"^Úroveň motivačního procesu jako způsobu vedení lidí$",
156
				"^The level of motivation process as a leadership$",
157
				"^Pay-beds in N(\\.?)H(\\.?)S(\\.?) Hospitals$",
158
				"(?i)^news and events$",
159
				"(?i)^NOVOSTI I DOGAĐAJI$",
160
				"^Sansū no gakushū$",
161
				"^Posouzení informačního systému firmy a návrh změn$",
162
				"^Information System Assessment and Proposal for ICT Modification$",
163
				"^Stresové zatížení pracovníků ve vybrané profesi$",
164
				"^Stress load in a specific job$",
165

    
166
				"^Sunday: Poster Sessions, Pt.*$",
167
				"^Monday: Poster Sessions, Pt.*$",
168
				"^Wednesday: Poster Sessions, Pt.*",
169
				"^Tuesday: Poster Sessions, Pt.*$",
170

    
171
				"^Analýza reklamy$",
172
				"^Analysis of advertising$",
173

    
174
				"^Shōgaku shūshinsho$",
175
				"^Shōgaku sansū$",
176
				"^Shintei joshi kokubun$",
177
				"^Taishō joshi kokubun dokuhon$",
178
				"^Joshi kokubun$",
179

    
180
				"^Účetní uzávěrka a účetní závěrka v ČR$",
181
				"(?i)^The \"?Causes\"? of Cancer$",
182
				"^Normas para la publicación de artículos$",
183
				"^Editor('|s)(s|') [Rr]eply$",
184
				"^Editor(’|s)(s|’) letter$",
185
				"^Redaktoriaus žodis$",
186
				"^DISCUSSION ON THE PRECEDING PAPER$",
187
				"^Kōtō shōgaku shūshinsho jidōyō$",
188
				"^Shōgaku nihon rekishi$",
189
				"^(Theory of the flow of action currents in isolated myelinated nerve fibers).*$",
190
				"^Préface$",
191
				"^Occupational [Hh]ealth [Ss]ervices.$",
192
				"^In Memoriam Professor Toshiyuki TAKESHIMA$",
193
				"^Účetní závěrka ve vybraném podniku.*$",
194
				"^Financial statements in selected company$",
195
				"^Abdominal [Aa]ortic [Aa]neurysms.*$",
196
				"^Pseudomyxoma peritonei$",
197
				"^Kazalo autora$",
198

    
199
				"(?i)^uvodna riječ$",
200
				"^Motivace jako způsob vedení lidí$",
201
				"^Motivation as a leadership$",
202
				"^Polyfunkční dům$",
203
				"^Multi\\-funkcional building$",
204
				"^Podnikatelský plán$",
205
				"(?i)^Podnikatelský záměr$",
206
				"(?i)^Business Plan$",
207
				"^Oceňování nemovitostí$",
208
				"^Marketingová komunikace$",
209
				"^Marketing communication$",
210
				"^Sumario Analítico$",
211
				"^Riječ uredništva$",
212
				"^Savjetovanja i priredbe$",
213
				"^Índice$",
214
				"^(Starobosanski nadpisi).*$",
215
				"^Vzdělávání pracovníků v organizaci$",
216
				"^Staff training in organization$",
217
				"^(Life Histories of North American Geometridae).*$",
218
				"^Strategická analýza podniku$",
219
				"^Strategic Analysis of an Enterprise$",
220
				"^Sadržaj$",
221
				"^Upute suradnicima$",
222
				"^Rodinný dům$",
223
				"(?i)^Fami(l)?ly house$",
224
				"^Upute autorima$",
225
				"^Strategic Analysis$",
226
				"^Finanční analýza vybraného podniku$",
227
				"^Finanční analýza$",
228
				"^Riječ urednika$",
229
				"(?i)^Content(s?)$",
230
				"(?i)^Inhalt$",
231
				"^Jinjō shōgaku shūshinsho jidōyō$",
232
				"(?i)^Index$",
233
				"^Chūgaku kokubun kyōkasho$",
234
				"^Retrato de una mujer$",
235
				"^Retrato de un hombre$",
236
				"^Kōtō shōgaku dokuhon$",
237
				"^Shotōka kokugo$",
238
				"^Shōgaku dokuhon$",
239
				"^Jinjō shōgaku kokugo dokuhon$",
240
				"^Shinsei kokugo dokuhon$",
241
				"^Teikoku dokuhon$",
242
				"^Instructions to Authors$",
243
				"^KİTAP TAHLİLİ$",
244
				"^PRZEGLĄD PIŚMIENNICTWA$",
245
				"(?i)^Presentación$",
246
				"^İçindekiler$",
247
				"(?i)^Tabl?e of contents$",
248
				"^(CODICE DEL BEATO DE LOS REYES FERNANDO I Y SANCHA).*$",
249
				"^(\\[MADRID\\. BIBL\\. NAC\\. N.*KING FERDINAND I.*FROM SAN ISIDORO DE LEON\\. FACUNDUS SCRIPSIT DATED.*\\]).*",
250
				"^Editorial( Board)?$",
251
				"(?i)^Editorial \\(English\\)$",
252
				"^Editörden$",
253
				"^(Corpus Oral Dialectal \\(COD\\)\\.).*$",
254
				"^(Kiri Karl Morgensternile).*$",
255
				"^(\\[Eksliibris Aleksandr).*\\]$",
256
				"^(\\[Eksliibris Aleksandr).*$",
257
				"^(Eksliibris Aleksandr).*$",
258
				"^(Kiri A\\. de Vignolles).*$",
259
				"^(2 kirja Karl Morgensternile).*$",
260
				"^(Pirita kloostri idaosa arheoloogilised).*$",
261
				"^(Kiri tundmatule).*$",
262
				"^(Kiri Jenaer Allgemeine Literaturzeitung toimetusele).*$",
263
				"^(Eksliibris Nikolai Birukovile).*$",
264
				"^(Eksliibris Nikolai Issakovile).*$",
265
				"^(WHP Cruise Summary Information of section).*$",
266
				"^(Measurement of the top quark\\-pair production cross section with ATLAS in pp collisions at).*$",
267
				"^(Measurement of the spin\\-dependent structure function).*",
268
				"(?i)^.*authors['’′]? reply\.?$",
269
				"(?i)^.*authors['’′]? response\.?$"
270
			]
271
		}
272
	}
273
}
(5-5/6)