Ngingathanda ukwethula emphakathini isiqeshana sale ncwadi esanda kushicilelwa:
Imodeli ye-Ontological yebhizinisi: izindlela nobuchwepheshe [Umbhalo]: i-monograph / [S. V. Gorshkov, S. S. Kralin, O. I. Mushtak nabanye; umhleli omkhulu S.V. Gorshkov]. - Ekaterinburg: I-Ural University Publishing House, 2019. - 234 p.: Ukugula, ithebula; 20 cm. - Umbhali. kukhonjiswe esinqeni sangemuva. Nge. - I-Bibliography ekupheleni kwech. - I-ISBN 978-5-7996-2580-1: amakhophi angu-200.
Inhloso yokuthumela lesi siqeshana ku-Habré iphindwe kane:
- Mancane amathuba okuthi kube khona ozokwazi ukubamba le ncwadi ngezandla uma engelona ikhasimende lomuntu ohlonishwayo ; Ayithengiswa neze.
- Ukulungiswa kwenziwe embhalweni (akugqanyiswanga ngezansi) futhi kwenziwe izengezo ezingahambelani kakhulu nefomethi ye-monograph ephrintiwe: amanothi ezihloko (ngaphansi kwama-spoilers) kanye nezixhumanisi.
- Ngiyafuna qoqa imibuzo namazwana, ukuze uzicabangele lapho uhlanganisa lo mbhalo ngendlela ebuyekeziwe kunoma yiziphi ezinye izincwadi.
- Abalandeli abaningi be-Semantic Web kanye Nedatha Exhunyiwe basakholelwa ukuthi umbuthano wabo mncane kakhulu, ikakhulukazi ngoba umphakathi jikelele awukakachazelwa kahle ukuthi kuhle kangakanani ukuba umlandeli weWebhu Ye-Semantic kanye Nedatha Exhunyiwe. Umbhali wocezu, nakuba engowalo mbuthano, akanawo lo mbono, kodwa, nokho, uzibona enesibopho sokwenza omunye umzamo.
Ngakho-ke,
I-Web Semantic
Ukuvela kwe-inthanethi kungamelwa kanje (noma sikhulume ngezigaba zayo ezakhiwe ngendlela eboniswe ngezansi):
- Amadokhumenti aku-inthanethi. Ubuchwepheshe obubalulekile - i-Gopher, i-FTP, njll.
I-inthanethi iyinethiwekhi yomhlaba wonke yokushintshisana ngezinsiza zendawo. - Imibhalo ye-inthanethi. Ubuchwepheshe obubalulekile yi-HTML ne-HTTP.
Imvelo yezinsiza eziveziwe ibheka izici ze-transmission medium yazo. - Idatha ye-inthanethi. Ubuchwepheshe obubalulekile - i-REST ne-SOAP API, i-XHR, njll.
Inkathi yezicelo ze-inthanethi, hhayi abantu kuphela ababa abathengi bezinsiza. - Idatha ye-inthanethi. Ubuchwepheshe obubalulekile ubuchwephesha Bedatha Exhunyiwe.
Lesi sigaba sesine, esabikezelwa nguBerners-Lee, umdali wobuchwepheshe besibili obuyisisekelo kanye nomqondisi we-W3C, sibizwa ngokuthi i-Semantic Web; Ubuchwepheshe Bedatha Exhunyiwe buklanyelwe ukwenza idatha kuwebhu ingafundeki umshini kuphela, kodwa futhi “iqondakala umshini.”
Kulokhu okulandelayo, umfundi uzoqonda ukuxhumana phakathi kwemiqondo ebalulekile yesigaba sesibili nesesine:
- Ama-URL ayafana nama-URI,
- i-analogue ye-HTML yi-RDF,
- Izixhumanisi ze-HTML zifana nezenzeko ze-URI kumadokhumenti e-RDF.
Iwebhu ye-Semantic ingaphezu kombono ohlelekile wekusasa le-inthanethi kunenkambiso ethile ezenzakalelayo noma ephoqelelwe, nakuba ingacabangela lokhu kwakamuva. Isibonelo, isici esibalulekile salokho okubizwa nge-Web 2.0 kuthathwa ngokuthi "okuqukethwe okukhiqizwa umsebenzisi." Ikakhulukazi, isincomo se-W3C siyacelwa ukuthi sicabangele ""kanye nomsebenzi onjalo .
Ingabe I-Semantic Web Ifile?
Uma wenqaba , isimo sewebhu ye-semantic sicishe sifane nesobukhomanisi ngezikhathi ze-socialism ethuthukisiwe (futhi vumela wonke umuntu azinqumele ukuthi ingabe ukwethembeka emithethweni enemibandela ka-Ilyich kuyaqashelwa). Izinjini zokusesha phoqa amawebhusayithi ukuthi asebenzise i-RFa ne-JSON-LD futhi ngokwawo asebenzise ubuchwepheshe obuhlobene nalokhu okuchazwe ngezansi (Igrafu Yolwazi Lwe-Google, Igrafu Yolwazi lwe-Bing).
Ngokuvamile, umbhali akakwazi ukusho ukuthi yini evimbela ukusabalala okukhulu, kodwa angakhuluma ngesisekelo sokuhlangenwe nakho kwakhe siqu. Kunezinkinga ezingaxazululwa "ngaphandle kwebhokisi" ezimweni zokuhlasela kwe-SW, nakuba zingandile kakhulu. Ngenxa yalokho, labo ababhekene nale misebenzi abanazo izindlela zokuphoqelela kulabo abakwazi ukunikeza isisombululo, kuyilapho ukuhlinzekwa okuzimele kwesixazululo sokugcina kuphikisana namamodeli abo ebhizinisi. Ngakho-ke siyaqhubeka nokuhlaziya i-HTML futhi sinamathisele ndawonye ama-API ahlukahlukene, enye nenye i-shittier.
Kodwa-ke, ubuchwepheshe beDatha Exhunyiwe buye basakaza ngale kweWebhu evamile; Incwadi, empeleni, inikezelwe kulezi zinhlelo zokusebenza. Okwamanje, umphakathi Wedatha Exhunyiwe ulindele ukuthi lobu buchwepheshe busakazeke kakhulu ngenxa yokuqoshwa kuka-Gartner (noma isimemezelo, ngendlela othanda ngayo) yamathrendi afanayo Amagrafu Olwazi и Indwangu Yedatha. Ngingathanda ukukholelwa ukuthi ngeke kube ukuqaliswa “kwebhayisikili” kwale miqondo okuzophumelela, kodwa leyo ehlobene namazinga e-W3C okukhulunywe ngawo ngezansi.
Idatha Exhunyiwe
UBerners-Lee uchaze Idatha Exhunyiwe njengewebhu ye-semantic "yenziwe kahle": isethi yezindlela nobuchwepheshe obuvumela ukuthi ifinyelele imigomo yayo yokugcina. Izimiso eziyisisekelo ze-Linked Data Berners-Lee Okulandelayo.
Isimiso 1. Ukusebenzisa ama-URI ukuqamba amabhizinisi.
Ama-URI ayizihlonzi zebhizinisi zomhlaba wonke ngokuphambene nezihlonzi zeyunithi yezinhlamvu zasendaweni zokufakiwe. Kamuva, lesi simiso sivezwe kahle kakhulu esiqubulweni se-Google Knowledge Graph “".
Isimiso 2. Ukusebenzisa ama-URI kuhlelo lwe-HTTP ukuze kususwe ireferensi.
Ngokubhekisela ku-URI, kufanele kwenzeke ukuthola okuphawuliwe ngemuva kwalowo mbonisi (isifaniso negama lesisebenzisi " sicacile lapha).*"ku-C); ngokunembe kakhudlwana, ukuze uthole ukumelwa okuthile kwalokhu - kuye ngevelu yesihloko se-HTTP Accept:. Mhlawumbe, ngokufika kwenkathi ye-AR/VR, kuzokwazi ukuthola insiza ngokwayo, kodwa okwamanje, cishe, kuzoba idokhumenti ye-RDF, okuwumphumela wokwenza umbuzo we-SPARQL. DESCRIBE.
Isimiso 3. Ukusetshenziswa kwamazinga e-W3C - ngokuyinhloko i-RDF(S) kanye ne-SPARQL - ikakhulukazi lapho kuhoxiswa ama-URI.
Lezi “zendlalelo” ezingazodwana zesitaki sobuchwepheshe be-Linked Data, esikwaziwa nangokuthi , izochazwa ngezansi.
Isimiso 4. Ukusetshenziswa kwezinkomba eziya kwamanye ama-URI uma kuchazwa amabhizinisi.
I-RDF ikuvumela ukuba ukhawule encazelweni yomlomo yensiza ngolimi lwemvelo, futhi umgomo wesine ubiza ukuthi ungakwenzi lokhu. Uma isimiso sokuqala sibhekwa emhlabeni wonke, siyenzeka lapho kuchazwa insiza ukubhekisela kwabanye, kuhlanganise "nabangaphandle", yingakho idatha ibizwa ngokuthi i-linked. Eqinisweni, cishe akunakugwenywa ukusebenzisa ama-URI aqanjwe kusilulumagama se-RDFS.
I-RDF
(Uhlaka Lwencazelo Yensiza) luwumthetho wokuchaza amabhizinisi ahlobene.
Izitatimende zohlobo lwe-“subject-predicate-object”, ezibizwa ngama-triplets, zenziwa mayelana nezinhlangano kanye nobudlelwano bawo. Esimeni esilula, isihloko, isilandiso, kanye nento wonke ama-URI. I-URI efanayo ingaba ezikhundleni ezihlukene ngamathathu ahlukene: ibe isihloko, isilandiso, kanye nento; Ngakho-ke, ama-triplets akha uhlobo lwegrafu olubizwa ngokuthi igrafu ye-RDF.
Izihloko kanye izinto kungaba hhayi URIs kuphela, kodwa futhi okuthiwa amanodi angenalutho, futhi izinto zingaba angokoqobo. Literals yizibonelo zezinhlobo zakudala ezihlanganisa ukumelwa kweyunithi yezinhlamvu kanye nenkomba yohlobo.
Izibonelo zokubhala amagama (ku-Turtle syntax, okuningi ngayo ngezansi): "5.0"^^xsd:float и "five"^^xsd:string. Literals ezinohlobo rdf:langString ingabuye ifakwe ithegi yolimi; ngoFudu kubhalwe kanje: "five"@en и "пять"@ru.
Amanodi angenalutho ayizinsiza “ezingaziwa” ezingenazo izihlonzi zomhlaba wonke, mayelana nokuthi yiziphi izitatimende, nokho, ezingenziwa; uhlobo oluguquguqukayo olukhona.
Ngakho (lokhu, empeleni, iphuzu lonke le-RDF):
- isihloko siyi-URI noma indawo engenalutho,
- isilandiso yi-URI,
- into iyi-URI, i-node engenalutho, noma ingokoqobo.
Kungani izilandiso zingakwazi ukuba amanodi angenalutho?
Isizathu esingaba khona isifiso sokuqonda ngokungakahleleki nokuhumusha i-triplet olimini lwe-oda lokuqala le-predicate logic. s p o njengento efana 
kuphi - isilandiso, и - ama-constants. Imikhondo yalokhu kuqonda ikudokhumenti “", enesimo senothi leqembu elisebenzayo le-W3C. Ngalokhu kuqonda, i-triplet s p []kuphi [] - i-node engenalutho, izohunyushwa ngokuthi kuphi - okuguquguqukayo, kodwa kanjani-ke ukuhumusha s [] o? Idokhumenti enesimo seSincomo se-W3C "” inikeza enye indlela yokuhumusha, kodwa namanje ayicabangi ukuthi kungenzeka ukuthi izilandiso zibe amanodi angenalutho.
Kodwa-ke, iManu Sporni .
I-RDF iyimodeli engabonakali. I-RDF ingabhalwa (i-serialized) kuma-syntaxes ahlukahlukene: , (efundeka kakhulu abantu), , (kanambambili).
I-RDF efanayo ingafakwa ku-RDF/XML ngezindlela ezahlukene, ngakho-ke, ngokwesibonelo, akwenzi mqondo ukuqinisekisa i-XML ewumphumela usebenzisa i-XSD noma uzame ukukhipha idatha usebenzisa i-XPath. Ngokunjalo, i-JSON-LD mancane amathuba okuthi yanelise isifiso sonjiniyela we-Javascript esimaphakathi sokusebenza ne-RDF isebenzisa ichashazi le-Javascript kanye nenothi lobakaki wesikwele (yize i-JSON-LD iya kuleyo ndawo ngokunikeza indlela ).
Iningi lama-syntaxes linikeza izindlela zokufinyeza ama-URI amade. Isibonelo, isikhangiso @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> kuFudu kuzobe sekukuvumela ukuthi ubhale esikhundleni salokho <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> nje rdf:type.
RDFS
(I-RDF Schema) - silulumagama esiyisisekelo sokumodela, sethula imiqondo yezakhiwo nesigaba nezakhiwo ezifana rdf:type, rdfs:subClassOf, rdfs:domain и rdfs:range. Kusetshenziswa isichazamazwi se-RDFS, isibonelo, izinkulumo ezilandelayo ezivumelekile zingabhalwa:
rdf:type rdf:type rdf:Property .
rdf:Property rdf:type rdfs:Class .
rdfs:Class rdfs:subClassOf rdfs:Resource .
rdfs:subClassOf rdfs:domain rdfs:Class .
rdfs:domain rdfs:domain rdf:Property .
rdfs:domain rdfs:range rdfs:Class .
rdfs:label rdfs:range rdfs:Literal .I-RDFS iyincazelo kanye nesilulumagama sokumodela, kodwa akulona ulimi oluyisithiyo (yize ukucaciswa okusemthethweni kanye ukusetshenziswa okunjalo). Igama elithi "Schema" akufanele liqondwe ngendlela efanayo naleyo eshiwo "i-XML Schema". Ngokwesibonelo, :author rdfs:range foaf:Person kusho lokho rdf:type wonke amanani ezakhiwo :author - foaf:Person, kodwa akusho ukuthi lokhu kufanele kushiwo kusenesikhathi.
I-SPARQL
(SPARQL Protocol and RDF Query Language) - ulimi lokubuza idatha ye-RDF. Esimeni esilula, umbuzo we-SPARQL isethi yamasampuli lapho ama-triplets amathathu egrafu ebuzwayo afaniswa nawo. Amaphethini angaqukatha okuguquguqukayo kusihloko, isilandiso, nendawo yento.
Umbuzo uzobuyisela amanani aguquguqukayo okuthi, uma eshintshwa kumasampula, aphumele esihlokweni segrafu ebuzwayo ye-RDF (isethi yama-triplets ayo). Okuguquguqukayo kwegama elifanayo kumasampuli ahlukene wama-triplets kufanele kube namanani afanayo.
Isibonelo, uma kunikezwe isethi engenhla yama-axiom ayisikhombisa e-RDFS, umbuzo olandelayo uzobuya rdfs:domain и rdfs:range njengamagugu ?s и ?p ngokulandelana:
SELECT * WHERE {
?s ?p rdfs:Class .
?p ?p rdf:Property .
}Kuyaphawuleka ukuthi i-SPARQL iyasho futhi ayilona ulimi lokuchaza ukuphenduleka kwegrafu (nokho, ezinye izinqolobane ze-RDF zinikeza izindlela zokulungisa uhlelo lokwenziwa kombuzo). Ngakho-ke, ezinye izinkinga zegrafu ezijwayelekile, ngokwesibonelo, ukuthola indlela emfushane kakhulu, azikwazi ukuxazululwa ku-SPARQL, okuhlanganisa ukusebenzisa (kodwa, futhi, amakhosombe e-RDF ngamanye anikeza izandiso ezikhethekile zokuxazulula lezi zinkinga).
I-SPARQL ayihlanganyeli ekucabangeni kokuvuleka komhlaba futhi ilandela indlela "yokuphika njengokwehluleka", lapho imiklamo efana FILTER NOT EXISTS {…}. Ukusatshalaliswa kwedatha kucatshangelwa kusetshenziswa indlela .
Indawo yokufinyelela ye-SPARQL - isitoreji se-RDF esikwazi ukucubungula imibuzo ye-SPARQL - ayinazo ama-analogue aqondile ukusuka esigabeni sesibili (bona ekuqaleni kwalesi sigaba). Ingafaniswa nedathabheyisi, ngokusekelwe kokuqukethwe lapho amakhasi e-HTML akhiqizwa khona, kodwa afinyeleleka ngaphandle. Iphoyinti lokufinyelela le-SPARQL lifana kakhulu nendawo yokufinyelela ye-API ukusuka esigabeni sesithathu, kodwa kunomehluko omkhulu emibili. Okokuqala, kungenzeka ukuhlanganisa imibuzo embalwa “ye-athomu” ibe yinto eyodwa (okubhekwa njengesici esiyinhloko se-GraphQL), futhi okwesibili, i-API enjalo iyazibhalela yona ngokuphelele (okuyikho i-HATEOAS eyazama ukuyifinyelela).
Ukuphawula kwepolitiki
I-RDF iyindlela yokushicilela idatha kuwebhu, ngakho isitoreji se-RDF kufanele sithathwe njenge-DBMS yedokhumenti. Kuyiqiniso, njengoba i-RDF iyigrafu hhayi isihlahla, nayo ivele yasuselwa kugrafu. Kuyamangaza ukuthi kuphumelele nhlobo. Ubani obengacabanga ukuthi kuzoba nabantu abahlakaniphile abazosebenzisa ama-node angenalutho. I-Codd lapha .
Kukhona nezindlela ezingafakiwe ngokugcwele zokuhlela ukufinyelela kudatha ye-RDF, isibonelo, (LDF) kanye (LDP).
OWL
(Ulimi lwe-Web Ontology) - okusemthethweni kokumela ulwazi, inguqulo ye-syntactic ye-logic yencazelo (yonke indawo ngezansi kulungile ukusho ukuthi I-OWL 2, inguqulo yokuqala ye-OWL yayisekelwe kuyo ).
Imiqondo yamagama achazayo ku-OWL ihambisana nezigaba, izindima zihambisana nezakhiwo, abantu ngabanye bagcina igama labo langaphambilini. Ama-axiom abizwa nangokuthi ama-axiom.
Ngokwesibonelo, kulokho okubizwa ngokuthi nge-OWL notation i-axiom esivele yaziwa kithi izobhalwa kanje:
Class: Human
Class: Parent
EquivalentClass: Human and (inverse hasParent) some Human
ObjectProperty: hasParentKunamanye ama-syntaxes okubhala i-OWL, njenge , esetshenziswa emibhalweni esemthethweni, kanye . Ukwengeza, i-OWL ingafakwa ku-serialized futhi ngokuqhubekayo - kunoma imaphi ama-syntaxes athile.
ISIKHOVA sinobudlelwano obubili ne-RDF. Ngakolunye uhlangothi, kungabhekwa njengohlobo lwesichazamazwi olunweba i-RDFS. Ngakolunye uhlangothi, kuwumthetho onamandla kakhulu lapho i-RDF ifomethi yochungechunge nje. Akuwona wonke ukwakheka kwe-OWL okuyisisekelo okungabhalwa kusetshenziswa i-RDF eyodwa kathathu.
Kuye ngokuthi iyiphi i-subset ye-OWL constructs evunyelwe ukusetshenziswa, ikhuluma ngalokho okubizwa ngokuthi . Amazinga ajwayelekile futhi adume kakhulu yi-OWL EL, OWL RL kanye ne-OWL QL. Ukukhethwa kwephrofayili kuthinta inkimbinkimbi yokubala yezinkinga ezijwayelekile. Isethi ephelele ye-OWL constructs ehambisanayo , ebizwa nge-OWL DL. Kwesinye isikhathi baphinde bakhulume nge-OWL Egcwele, lapho ukwakhiwa kwe-OWL kuvunyelwe ukusetshenziswa ngenkululeko egcwele etholakala ku-RDF, ngaphandle kwemikhawulo ye-semantic neyokubala. . Isibonelo, okuthile kungaba kokubili isigaba kanye nempahla. I-OWL Full ayinakunqunywa.
Izimiso ezibalulekile zokunamathisela imiphumela ku-OWL ukwamukelwa kokuqagela komhlaba ovulekile. ) kanye nokwenqatshwa kokuqagelwa kwamagama ahlukile (ukuqagela kwegama eliyingqayizivele, ). Ngezansi sizobona ukuthi lezi zimiso zingaholela kuphi futhi sethule ezinye izakhiwo ze-OWL.
Vumela i-ontology iqukathe lesi siqeshana esilandelayo (ku-syntax yaseManchester):
Class: manyChildren
EquivalentTo: Human that hasChild min 3
Individual: John
Types: Human
Facts: hasChild Alice, hasChild Bob, hasChild CarolIngabe kuzolandela kulokho okushiwo ukuthi uJohn unezingane eziningi? Ukwenqaba i-UNA kuzophoqa injini yemibono ukuthi iphendule lo mbuzo ngenegethivu, njengoba u-Alice no-Bob bangase babe umuntu oyedwa. Ukuze lokhu okulandelayo kwenzeke, uzodinga ukwengeza i-axiom elandelayo:
DifferentIndividuals: Alice, Bob, Carol, JohnManje ake ucezu lwe-ontology lube naleli fomu elilandelayo (uJohn kuthiwa unezingane eziningi, kodwa unezingane ezimbili kuphela):
Class: manyChildren
EquivalentTo: Human that hasChild min 3
Individual: John
Types: Human, manyChildren
Facts: hasChild Alice, hasChild Bob
DifferentIndividuals: Alice, Bob, Carol, JohnIngabe le ontology izobe ingahambisani (okungahunyushwa njengobufakazi bedatha engavumelekile)? Ukwamukela i-OWA kuzobangela injini yokukhomba ukuthi iphendule ngendlela engalungile: "kwenye indawo" (kwenye i-ontology) kungase kushiwo ukuthi uCarol naye uyingane kaJohn.
Ukukhipha ukuthi kungenzeka lokhu, ake sengeze iqiniso elisha ngoJohane:
Individual: John
Facts: hasChild Alice, hasChild Bob, not hasChild CarolUkuze ungafaki ukubukeka kwezinye izingane, ake sithi wonke amanani entengo "ukuba nengane" angabantu, esinawo amane kuphela:
ObjectProperty: hasChild
Domain: Human
Сharacteristics: Irreflexive
Class: Human
EquivalentTo: { Alice, Bill, Carol, John }Manje i-ontology izobe isiphikisana, okuyinto injini engeke yehluleke ukuyibika. Ngama-axioms wokugcina, ngomqondo othile, “sivale” umhlaba, futhi phawula ukuthi kungenzeka kanjani ukuthi uJohane abe ingane yakhe akhishwe.
Ixhumanisa Idatha Yebhizinisi
Isethi Yedatha Exhunyiwe yezindlela nobuchwepheshe ekuqaleni yayihloselwe ukushicilela idatha kuwebhu. Ukusetshenziswa kwabo endaweni yebhizinisi yangaphakathi kubhekene nenani lobunzima.
Isibonelo, endaweni evaliwe yebhizinisi, amandla okudonswa kwe-OWL asekelwe ekwamukelweni kwe-OWA kanye nokwenqatshwa kwe-UNA, izinqumo ngenxa yemvelo evulekile nesatshalaliswa yeWebhu, abuthakathaka kakhulu. Futhi lapha izixazululo ezilandelayo zingenzeka.
- Ukunikeza i-OWL nge-semantics, okusho ukushiywa kwe-OWA kanye nokwamukelwa kwe-UNA, ukuqaliswa kwenjini ehambisanayo yokuphumayo. - Ngale ndlela Isitoreji se-Stardog RDF.
- Ilahla amandla e-OWL okudonsela phansi ukuze ivune izinjini zokubusa. - I-Stardog isekela ; Ukunikezwa kwe-Jena ne-GraphDB imithetho
- Ukwenqatshwa kwamandla okudonswa kwe-OWL, ukusetshenziswa kwesethi engaphansi eyodwa noma enye eduze ne-RDFS ukuze kwenziwe imodeli. - Bona okwengeziwe ngalokhu ngezansi.
Enye inkinga ukugxila okukhulu izwe lebhizinisi okungenzeka libe nakho ezindabeni zekhwalithi yedatha kanye nokuntuleka kwamathuluzi okuqinisekisa idatha kusitaki Sedatha Exhunyiwe. Imiphumela lapha imi kanje.
- Futhi, sebenzisa ukuze kuqinisekiswe ukwakhiwa kwe-OWL ene-semantics yomhlaba evaliwe namagama ahlukile uma injini efanelekile yokukhomba itholakala.
- Sebenzisa , elimisiwe ngemva kokuba uhlu lwezendlalelo ze-Semantic Web Layer Cake selulungisiwe (nokho, lungasetshenziswa futhi njengenjini yemithetho), noma .
- Ukuqonda ukuthi yonke into ekugcineni yenziwa ngemibuzo ye-SPARQL, kwakheka eyakho indlela elula yokuqinisekisa idatha usebenzisa yona.
Nokho, ngisho nokwenqatshwa okuphelele kwamandla emali namathuluzi okuqinisekisa kushiya isitaki Sedatha Exhunyiwe ngaphandle kokuncintisana emisebenzini efana nokwakheka kwezwe kuwebhu evuliwe nesabalalisiwe - emisebenzini yokuhlanganisa idatha.
Kuthiwani ngesistimu yolwazi lwebhizinisi evamile?
Lokhu kungenzeka, kodwa kufanele, vele, uqaphele ukuthi yiziphi izinkinga ubuchwepheshe obuhambisanayo okufanele buzixazulule. Ngizochaza lapha ukusabela okujwayelekile kwabahlanganyeli bentuthuko ukukhombisa ukuthi lesi sitaki sobuchwepheshe sibukeka kanjani ngokombono we-IT evamile. Ingikhumbuza kancane ngomfanekiso wendlovu:
- Umhlaziyi webhizinisi: I-RDF into efana nemodeli enengqondo egcinwe ngokuqondile.
- Umhlaziyi Wezinhlelo: I-RDF ifana , kuphela ngenqwaba yezinkomba kanye nolimi lombuzo olulula.
- Unjiniyela: kahle, konke lokhu kusemoyeni wemiqondo yemodeli ecebile nekhodi ephansi, muva nje ngalokhu.
- Umphathi wephrojekthi: yebo kuyafana !
Ukuzijwayeza kubonisa ukuthi isitaki sivame ukusetshenziswa kakhulu emisebenzini ehlobene nokusatshalaliswa nokuhlukahluka kwedatha, isibonelo, lapho kwakhiwe amasistimu ekilasi e-MDM (Master Data Management) noma i-DWH (Data Warehouse). Izinkinga ezinjalo zikhona kunoma iyiphi imboni.
Ngokuya ngezinhlelo zokusebenza eziqondene nemboni, ubuchwepheshe bedatha obuxhumene budume kakhulu kulezi zimboni ezilandelayo.
- ubuchwepheshe be-biomedical (lapho ukuthandwa kwabo kubonakala kuhlobene nobunkimbinkimbi besizinda);
okwamanje
“I-Boiling Point” isanda kusingatha ingqungquthela ehlelwe yinhlangano “yeNational Medical Knowledge Base” “".
- ukukhiqizwa nokusebenza kwemikhiqizo eyinkimbinkimbi (ubunjiniyela obukhulu bemishini, ukukhiqizwa kukawoyela negesi; imvamisa sikhuluma ngezinga elijwayelekile );
okwamanje
Lapha, futhi, isizathu siyinkimbinkimbi yendawo yesifundo, lapho, isibonelo, esiteji esikhuphukayo, uma sikhuluma ngemboni yamafutha negesi, ukubala okulula kudinga imisebenzi ethile ye-CAD.
Ngo-2008, kwenzeka umcimbi wokufaka omele, owawuhlelwe yi-Chevron .
I-ISO 15926, ekugcineni, yabonakala inzima kancane embonini kawoyela negesi (futhi yatholakala mhlawumbe ukusetshenziswa okukhulu kubunjiniyela bemishini). Kuphela iStatoil (i-Equinor) eyabanjwa ngokuphelele kuyo; eNorway, iyonke . Abanye bazama ukuzenzela okwabo. Isibonelo, ngokusho kwamahemuhemu, uMnyango Wezamandla wasekhaya uhlose ukudala "imodeli ye-ontological yomqondo we-fuel and energy complex," efanayo, ngokusobala, .
- izinhlangano zezezimali (ngisho ne-XBRL ingathathwa njengohlobo lwenhlanganisela ye-SDMX kanye ne-ontology ye-RDF Data Cube);
okwamanje
Ekuqaleni konyaka, i-LinkedIn yafaka ugaxekile kumbhali ngezikhala zezikhala cishe kuzo zonke iziqhwaga zemboni yezezimali, azi ochungechungeni lwe-TV "Force Majeure": Goldman Sachs, JPMorgan Chase kanye/noma uMorgan Stanley, Wells Fargo, I-SWIFT/Visa/Mastercard, Bank of America, Citigroup, Fed, Deutsche Bank... Cishe wonke umuntu wayefuna umuntu angamthumela kuye. . Abambalwa bakwazile ukuthola: izinhlangano zezezimali zithathe yonke into .
Ku-HeadHunter, i-Sberbank kuphela ethole okuthile okuthakazelisayo; kwakumayelana "nesitoreji se-EAV esinemodeli yedatha efana ne-RDF."
Mhlawumbe, umehluko ngezinga lothando lobuchwepheshe obuhambisanayo bezikhungo zezimali zasekhaya nezaseNtshonalanga kungenxa yemvelo yamazwe ngamazwe yemisebenzi yakamuva. Ngokusobala, ukuhlanganiswa kuyo yonke imingcele yezwe kudinga izixazululo ezihlukene ngokwekhwalithi zenhlangano nezobuchwepheshe.
- amasistimu wokuphendula imibuzo anezicelo zezentengiso (IBM Watson, Apple Siri, Google Knowledge Graph);
okwamanje
Phela, umdali we-Siri, uThomas Gruber, ungumbhali wayo kanye incazelo ye-ontology (ngomqondo we-IT) ngokuthi "incazelo yomqondo." Ngokubona kwami, ukuhlela kabusha amagama akule ncazelo akushintshi incazelo yawo, okungenzeka ukuthi ikhombisa ukuthi ayikho.
- ukushicilelwa kwedatha ehlelekile (ngokufaneleka okukhulu lokhu kungase kubalulwe Kudatha Evuliwe Exhunyiwe).
okwamanje
Abalandeli abakhulu bedatha exhunyiwe yilabo ababizwa ngokuthi i-GLAM: Amagalari, Amalabhulali, Izingobo Zomlando, kanye Namamnyuziyamu. Kwanele ukusho ukuthi i-Library of Congress ikhuthaza ukushintshwa kwe-MARC21 , okuyinto inikeza isisekelo sekusasa lencazelo yezincwadi futhi, kunjalo, ngokusekelwe ku-RDF.
I-Wikidata ivame ukushiwo njengesibonelo sephrojekthi eyimpumelelo emkhakheni we-Linked Open Data - uhlobo lwenguqulo ye-Wikipedia efundeka ngomshini, okuqukethwe kwayo, ngokungafani ne-DBPedia, kungakhiqizwe ngokungeniswa kwamanye amazwe kusuka kumabhokisi emininingwane, kodwa kudalwe okuningi noma okuncane ngesandla (bese kuba umthombo wolwazi wamabhokisi emininingwane afanayo).
Sincoma futhi ukuthi uyihlole abasebenzisi besitoreji se-Stardog RDF kuwebhusayithi ye-Stardog esigabeni "Amakhasimende".
Yiba noma kunjalo, eGartner "I-Enterprise Taxonomy and Ontology Management" ibekwe phakathi kokwehlela esigodini sokuphoxeka ngethemba lokufinyelela "ethafeni lokukhiqiza" ngaphambi kweminyaka eyi-10.
Ixhuma Idatha Yebhizinisi
Izibikezelo, izibikezelo, izibikezelo...
Ngenxa yentshisekelo yomlando, ngibhale ngezansi izibikezelo zikaGartner zeminyaka eyahlukene ngobuchwepheshe esibuthakaselayo.
| I-Год | Ubuchwepheshe | Bika | Isikhundla | Iminyaka ukuya ethafeni |
|---|---|---|---|---|
| 2001 | I-Web Semantic | I-Emerging Technologies | I-Innovation Trigger | 5-10 |
| 2006 | Iwebhu ye-Semantic yebhizinisi | I-Emerging Technologies | Inani Eliphakeme Lokulindelwe Okwenyukayo | 5-10 |
| 2012 | I-Web Semantic | Idatha enkulu | Inani Eliphakeme Lokulindelwe Okwenyukayo | > 10 |
| 2015 | Idatha Exhunyiwe | Izibalo Ezithuthukile Nesayensi Yedatha | Umkhumbi Wokudumazeka | 5-10 |
| 2016 | Ukuphathwa kwe-Enterprise Ontology | I-Emerging Technologies | Umkhumbi Wokudumazeka | > 10 |
| 2018 | Amagrafu Olwazi | I-Emerging Technologies | I-Innovation Trigger | 5-10 |
Nokho, kakade ku sekuvele omunye umkhuba obheke phezulu - Ulwazi Graphs. Ukuphindukuzalwa okuthile kwenzeka: igrafu DBMSs, lapho ukunakwa kwabasebenzisi kanye nemizamo yabathuthukisi kwashintsha, ngaphansi kwethonya lezicelo zangaphambili kanye nemikhuba yakamuva, yaqala ukuthatha ama-contours nokubeka indawo. yezimbangi zabo zangaphambili.
Cishe yonke igrafu i-DBMS manje isizimemezela iyinkundla efanelekile yokwakha “igrafu yolwazi” yebhizinisi (“idatha exhunyiwe” ngezinye izikhathi ithathelwa indawo “idatha exhunyiwe”), kodwa zithetheleleka kangakanani izimangalo ezinjalo?
Imininingo egciniwe yegrafu isasemantic; idatha ekugrafu ye-DBMS iseyisizinda sedatha esifanayo. Izihlonzi zezintambo esikhundleni sama-URI zenza umsebenzi wokuhlanganisa ama-DBMS egrafu amabili kusengumsebenzi wokuhlanganisa, kuyilapho ukuhlanganisa izitolo ezimbili ze-RDF kuvame ukwehlela ekuhlanganiseni amagrafu amabili e-RDF. Esinye isici se-asemanticity ukungaguquguquki kwemodeli yegrafu ye-LPG, okwenza kube nzima ukuphatha imethadatha usebenzisa inkundla efanayo.
Okokugcina, ama-DBMS egrafu awanazo izinjini zokukhomba noma izinjini zokubusa. Imiphumela yezinjini ezinjalo ingaphinda ikhiqizwe ngemibuzo enzima, kodwa lokhu kungenzeka ngisho naku-SQL.
Kodwa-ke, amasistimu okugcina e-RDF aholayo awanabo ubunzima bokusekela imodeli ye-LPG. Indlela eqine kakhulu yile ehlongozwa ngesikhathi esisodwa ku-Blazegraph: imodeli ye-RDF*, ehlanganisa i-RDF ne-LPG.
Funda kabanzi
Ungafunda kabanzi mayelana nokusekelwa kwesitoreji se-RDF semodeli ye-LPG esihlokweni esedlule se-Habré: . Ngithemba ukuthi ngolunye usuku kuzobhalwa i-athikili ehlukile mayelana ne-Knowledge Graphs ne-Data Fabric. Ingxenye yokugcina, njengoba kulula ukuyiqonda, yabhalwa ngokushesha, noma kunjalo, ngisho nezinyanga eziyisithupha kamuva, yonke into ayicacisi ngale mibono.
Izincwadi
- Halpin, H., Monnin, A. (abahleli) (2014). Ubunjiniyela Befilosofi: Ngaya Kwifilosofi Yewebhu
- Allemang, D., Hendler, J. (2011) Iwebhu yeSemantic ye-Working Ontologist (2nd ed.)
- Staab, S., Studer, R. (edits.) (2009) I-Handbook on Ontologies (2nd ed.)
- Wood, D. (ed.). (2011) Ukuxhumanisa Idatha Yebhizinisi
- Keet, M. (2018) Isingeniso Sobunjiniyela Be-Ontology
Source: www.habr.com