Uhlalutyo Cluster. Indlela yezenzululwazi izifundo ngezenzeko ezimbaxa

Ukulawulwa nayiphi na inkqubo, kuquka ukuthengisa, ibandakanya uvavanyo kwinjongo imeko emarikeni. Ngcembe lihambe inkqubo uhlalutyo yamathuba emakethe, eziquka ukhetho kweemarike ekujoliswe, kunye nophuhliso lorhwebo entsonkothileyo nozalisekiso lwemisebenzi yoshishino, wajamelana bengaqondanga kunye imfuneko yophando. Ngoko kuyimfuneko nje kuphela bathembele italente kunye namava umhlalutyi, kodwa uqeqesho ngokusebenzisa ubuchule lweenkcukacha.

Kuqoqosho lwale mihla, kunye neenkqubo zayo entsonkothileyo yaye ezininzi, izixa enkulu ulwazi ukufumana iinkcukacha ezininzi ezifanelekileyo ngaphandle kokusebenzisa iiphakheji ezahlukeneyo manani kuba yingxaki kakhulu.

Singene uhlalutyo ezisebenzisanayo indima ekhethekileyo kuphando sokuthengisa. Ngobume bayo, le ndlela zidibene, edibanisa iindlela eziliqela uphando lwezibalo. Oku kusekelwe ubuxoki zokuhlela nimana multivariate, nganye eneyazo isethi yaso variables esichazayo. Uhlalutyo Cluster lubonisa indlela Cazulula into (uniform) amaqela isalamane yohlobo olunye ukuba isethi lokuqala variables ukuze bayithathele ingqalelo. Ngamanye amazwi, izinto ziba ngamaqela. Ngokwamaqela, abonisa kwihi ngezizathu ezininzi.

iindlela Uhlalutyo Cluster zisetyenziselwa uluhlu olubanzi iinjongo zorhwebo.

segmentation Market kuvumela umsebenzisi ukuba ndimaphule kudidi ngokwamaqela ngokwesiseko iinzuzo ezilindelekileyo ekufunyanweni iimpahla ezithile. Iqela ngalinye kuba abasebenzisi abafuna ithuba iinzuzo ezifanayo. Igama athathe efanelekileyo - iingenelo indlela segmentation.

Uhlaziyo yokuziphatha komthengi. Kulo msebenzi, uhlalutyo iklasta isetyenziswa ukwenza i kumaqela yohlobo olunye umthengi ukuze akwazi ukuziphatha kwabo.

Ukuchaza iimpawu imveliso entsha, sinako ukuvelisa oko Kuphononongwa brand, kwangaxeshanye kufumaniseka ukuba aba njalo libizwe xa iibrand yesihloko umboniso elifanayo ukhuphiswano omnye nomnye ngaphandle kunye amanqaku namanye amaqela.

Ngokwamaqela izihloko phakathi komzi, ungakhetha iimarike zifanelekileyo iimpahla ezithile.

Uhlalutyo Cluster kunciphisa dimensionality lwe data. Yenza imigqaliselo izihloko ezahlukeneyo, uze udlulele uhlalutyo ezininzi discriminant. Kulula kakhulu kwaye zitshiphu ingqalelo kwimeko nganye.

Injongo wokunqumbana ukuba izinto iqela ngeempawu ezifanayo. Ukuze uvavanyo njongo ngakumbi iqondo ukufana kufuneka kuvelisa unit reference. Xa ukwakha izihloko ngokuqhelekileyo bathembele ezimbini okanye ngaphezulu phawu ngaxeshanye.

Uhlalutyo Cluster lubandakanya nokusetyenziswa koluhlu olubanzi iindlela yokunqumbana. Phakathi kwazo kukho ezifana indlela probabilistic, iindlela, ezo zisekelwe basemoyeni, indlela ingqiqo, indlela oluneqela elinegunya.

Uhlalutyo Iqela oluneqela ibandakanya inkqubo entsonkothileyo ukuba unalo inani lamaqela ekwiseli okanye ezisebenzisanayo odolo ezahlukeneyo. Le ndlela isebenzisa iintlobo ezimbini iimpawu. Agglomerate (ubunye) imiqondiso ahlalisane nama divizivnymi (ini). Inani iimpawu kukhokelela kwahlulwe ngeendlela yokuhlela monothetic kunye polythetic.

Ukusebenzisa zonke ezi ndlela e manani, kukho malunga ezilikhulu kanye ubuchule yokunqumbana. Kodwa analysis Iqela oluneqela ihamba phambili kolu luhlu. ukhanga usendleleni yokuba esebenza ngokupheleleyo kunye intswelo data, naxa idata ekhoyo akwenzeki njengoko kufunwa yi iimeko ngokuqhelekileyo zisasazwe eziguquguqukayo random, kwakunye nezinye iimfuno iindlela yamandulo manani.