Matematicki pristup socijalnom umrežavanju

U ovom radu promatrali smo stohastičke modele za društvene mreže podataka, te uveli posebne vrste statističkih distribucija za slučajno usmjerene grafove i pokazali da su poseban slučaj uniformne distribucije slučajno usmjerenog grafa. Prezentirali smo modele za mrežu s mjerenjima za jednostruke, direktne veze za jednu grupu promatranih podataka. Zatim smo opisali i prikazati interpretaciju i prilagodbu osnovnog statističkog modela za mrežu. Predstavili smo modele napravljene za proučavanje kako veze od slobodnih odnosa variraju preko članova. Ovdje opisani modeli su modeli dijadičke interakcije koji koriste (prirodni) logaritam vjerojatnosti kao svoju osnovnu mjeru.In this diploma thesis we considered a class of statistical distributions for random directed graphs, which, as we have shown, is a special case of the uniform random directed graph distributions. We have presented models for a network with measurements on a single, directional relation for one set of actors. We have then described and demonstrated the interpretation and fitting of a basic statistical network model. Also, we can study models that are focused only on ties between groups of actors or just single actor, We have presented a class of models designed to study how ties from a single relation vary across actors. The models described here are dyadic interaction models, which use the (natural) log of probabilities as their basic modeling unit