Ababhali abaninzi bangaphandle bayavuma ukuba ixesha leDatha enkulu lifikelele esiphelweni. Kwaye kule meko, igama elithi Big Data libhekiselele kubuchwephesha obusekwe kwiHadoop. Ababhali abaninzi banokuwubiza ngokuzithemba umhla apho iDatha enkulu ishiye emhlabeni kwaye lo mhla yi-05.06.2019/XNUMX/XNUMX.
Kwenzeka ntoni ngolu suku lubalulekileyo?
Ngolu suku, i-MAPR ithembise ngokuwunqumamisa umsebenzi wayo ukuba ayifumani mali yokuqhubeka nokusebenza. I-MAPR kamva yafunyanwa yi-HP ngo-Agasti ka-2019. Kodwa ukubuyela ngoJuni, umntu akanakunceda kodwa aqaphele intlekele yeli xesha kwimarike yeDatha enkulu. Le nyanga yabona ukuwa kwamaxabiso e-stock CLOUDERA, umdlali ohamba phambili kwimarike, edibaniswe ne-HORTOWORKS engapheliyo ngoJanuwari wonyaka ofanayo. Ukuwa kwakubaluleke kakhulu kwaye kwafikelela kwi-43%, ekugqibeleni, imali ye-CLOUDERA yehla ukusuka kwi-4,1 ukuya kwi-1,4 yeebhiliyoni zeedola.
Akunakwenzeka ukuba ungathethi ukuba amahemuhemu e-bubble kwintsimi ye-Hadoop-based teknoloji iye yajikeleza ukususela ngoDisemba 2014, kodwa ibambelele ngesibindi iminyaka ephantse ibe mihlanu. La mahemuhemu ayesekelwe ekwaleni kukaGoogle, inkampani apho iteknoloji yeHadoop yavela khona, ekuyilweni kwayo. Kodwa itekhnoloji ithathe iingcambu ngexesha lotshintsho lweenkampani ukuya kwizixhobo zokusebenza zelifu kunye nophuhliso olukhawulezayo lobukrelekrele bokwenziwa. Ngoko ke, xa sikhangela emva, sinokuthi ngentembelo ukufa kwakulindelwe.
Ngaloo ndlela, ixesha leDatha enkulu lifikelele esiphelweni, kodwa kwinkqubo yokusebenza kwiDatha enkulu, iinkampani ziye zaqaphela zonke iinqununu zokusebenza kuyo, izibonelelo ezinokuthi iDatha enkulu ikwazi ukuzisa ishishini, kwaye yafunda ukusebenzisa i-artificial. ubukrelekrele bokukhupha ixabiso kwidatha ekrwada.
Okona kunomdla ngakumbi kuba ngumbuzo wento eza kuthatha indawo yetekhnoloji kunye nendlela itekhnoloji yokuhlalutya buya kukhula ngakumbi.
Uhlalutyo olongeziweyo
Ngexesha leziganeko ezichazwe, iinkampani ezisebenza kwintsimi yohlalutyo lwedatha azizange zihlale. Yintoni enokugwetywa ngokusekelwe kulwazi malunga neentengiselwano ezenzeke ngo-2019. Kulo nyaka, ukuthengiselana okukhulu kwimarike kwenziwa - ukufunyanwa kweqonga lokuhlalutya i-Tableau yi-Salesforce ye-15,7 yeebhiliyoni zeedola. Isivumelwano esincinci senzeka phakathi kukaGoogle kunye noJonga. Kwaye ke, umntu akanako ukusilela ukuqaphela ukufunyanwa nguQlik weqonga elikhulu ledatha yeAttunity.
Iinkokeli zentengiso ye-BI kunye neengcali zikaGartner zibhengeza utshintsho olukhulu kwiindlela zokuhlalutya idatha; olu tshintsho luya kuyitshabalalisa ngokupheleleyo imarike ye-BI kwaye ikhokelele ekutshintshweni kwe-BI nge-AI. Kulo mongo, kufuneka kuqatshelwe ukuba isishunqulelo se-AI asiyiyo "ubukrelekrele bokwenziwa" kodwa "uBukrelekrele obongezelelweyo". Makhe sihlolisise oko kusemva kwamagama athi "Uhlalutyo olongezelelekileyo."
Uhlahlelo olongeziweyo, njengenyani eyongeziweyo, lusekwe kwiingxelo ezininzi ngokubanzi:
- ukukwazi ukunxibelelana usebenzisa i-NLP (i-Natural Language Processing), okt. ngolwimi lomntu;
- ukusetyenziswa kobukrelekrele bokwenziwa, oku kuthetha ukuba idatha iya kulungiswa kwangaphambili ngobukrelekrele bomatshini;
- kwaye ngokuqinisekileyo, iingcebiso ezifumanekayo kumsebenzisi wenkqubo, eziveliswe ngobukrelekrele bokwenziwa.
Ngokutsho kwabavelisi bamaqonga okuhlalutya, ukusetyenziswa kwabo kuya kufumaneka kubasebenzisi abangenazo izakhono ezikhethekileyo, njengolwazi lwe-SQL okanye ulwimi olufanayo lokubhala, abangenalo uqeqesho lwamanani okanye lwemathematika, abangenalo ulwazi lweelwimi ezidumileyo. ukugqwesa ukusetyenzwa kwedatha kunye namathala eencwadi ahambelanayo. Abantu abanjalo, ababizwa ngokuba “ziNzululwazi zeDatha yabemi”, kufuneka babe neziqinisekiso zoshishino ezibalaseleyo kuphela. Umsebenzi wabo kukubamba ulwazi lweshishini kwiingcebiso kunye noqikelelo oluya kubanika ubukrelekrele bokwenziwa, kwaye banokucokisa ukuqikelela kwabo besebenzisa i-NLP.
Ukuchaza inkqubo yabasebenzisi abasebenza kunye neenkqubo zale klasi, umntu unokucinga ngomfanekiso olandelayo. Umntu, oza emsebenzini aze aqalise isicelo esihambelanayo, ukongeza kwiseti eqhelekileyo yeengxelo kunye needashbhodi ezinokuthi zihlalutywe kusetyenziswa iindlela ezisemgangathweni (ukuhlelwa, ukwahlula, ukwenza imisebenzi ye-arithmetic), ubona iingcebiso ezithile kunye neengcebiso, into efana nale: ukuze ufezekise i-KPI, inani leentengiso, kufuneka ufake isaphulelo kwiimveliso ezivela kudidi “lweGardening”. Ukongeza, umntu unokuqhagamshelana nomthunywa wenkampani: Skype, Slack, njl. Unokubuza imibuzo yerobhothi, ngesicatshulwa okanye ngelizwi: "Ndinike abona bathengi bahlanu banengeniso." Emva kokuba efumene impendulo efanelekileyo, kufuneka enze esona sigqibo sifanelekileyo ngokusekelwe kumava akhe oshishino kwaye azise inzuzo kwinkampani.
Ukuba uthatha inyathelo emva kwaye ujonge ukubunjwa kolwazi oluhlalutywayo, kwaye ngeli nqanaba, iimveliso zokuhlalutya ezongeziweyo zinokwenza ubomi babantu bube lula. Ngokufanelekileyo, kucingelwa ukuba umsebenzisi uya kufuna kuphela ukukhomba imveliso yohlalutyo kwimithombo yolwazi olufunwayo, kwaye inkqubo ngokwayo iya kunyamekela ukudala imodeli yedatha, iitafile zokudibanisa kunye nemisebenzi efanayo.
Konke oku kufuneka, okokuqala, kuqinisekise "idemokhrasi" yedatha, okt. Nawuphi na umntu unokuhlalutya lonke uluhlu lolwazi olukhoyo kwinkampani. Inkqubo yokwenza izigqibo kufuneka ixhaswe ngeendlela zokuhlalutya izibalo. Ixesha lokufikelela kwidatha kufuneka libe lincinci, ngoko akukho mfuneko yokubhala izikripthi kunye nemibuzo ye-SQL. Kwaye kunjalo, unokonga imali kwiingcali zeSayensi yeDatha ehlawulwa kakhulu.
Ngokwenyani, itekhnoloji ibonelela ngamathemba aqaqambileyo kwishishini.
Yintoni ethatha indawo yeDatha enkulu?
Kodwa, enyanisweni, ndaqala inqaku lam ngeDatha enkulu. Kwaye andikwazanga ukuphuhlisa esi sihloko ngaphandle kohambo olufutshane kwizixhobo zanamhlanje ze-BI, isiseko esihlala sikwiDatha enkulu. Isiphelo sedatha enkulu ngoku sinqunywe ngokucacileyo, kwaye iteknoloji yefu. Ndigxininise kwizivumelwano ezenziwe nabathengisi be-BI ukwenzela ukubonisa ukuba ngoku yonke inkqubo yokuhlalutya inokugcinwa kwamafu emva kwayo, kwaye iinkonzo zefu zine-BI njengesiphelo sangaphambili.
Ngaphandle kokulibala malunga neentsika ezinjalo kwintsimi yogcino-lwazi njenge-ORACLE kunye neMicrosoft, kuyimfuneko ukuqaphela isikhokelo sabo esikhethiweyo sophuhliso lweshishini kwaye eli lifu. Zonke iinkonzo ezibonelelwayo zinokufumaneka efini, kodwa ezinye iinkonzo zelifu azisafumaneki kwindawo. Benze umsebenzi obalulekileyo ekusetyenzisweni kweemodeli zokufunda koomatshini, badale amathala eencwadi afumanekayo kubasebenzisi, kunye nojongano olucwangcisiweyo ukuze kube lula ukusebenza kunye neemodeli ukusuka ekukhetheni ukuseta ixesha lokuqala.
Enye inzuzo ebalulekileyo yokusebenzisa iinkonzo zefu, ezivakaliswa ngabavelisi, kukufumaneka kweeseti zedatha ezingapheliyo nakwesiphi na isihloko kwiimodeli zoqeqesho.
Nangona kunjalo, umbuzo uvela: iteknoloji yefu iya kuthatha ingcambu kangakanani kwilizwe lethu?
umthombo: www.habr.com