hrvatski

Moti u kompleksnim mrežama jezika

Complex networks are graphs with topological features which cannot be found in simple, random graphs. Many systems in nature can be described as complex networks, where objects are vertices and connections among those objects are edges. Those edges can be directed and undirected. In this thesis focus was on complex network of language as network of words connected with the edge where they follow one another. Motif is a subgraph of a graph, with specif c pattern of connections among vertices. This thesis researches weighted motifs (directed) with three vertices. The weight of edge is the frequency of paired words connected with that edge. There are thirteen possible types of motifs with three vertices. Calculations are made for number of possibilities of weight formation for each one of them. Knowledge of graph theory and combinatorics are used in forming those calculations. In this thesis there was an experiment ran with gtrieScanner software on 45 texts ; 15 in one of three languages: English, Italian and Croatian. Frequencies of motifs for each language were analyzed with the use of statistics. Type 2 motif is characteristic motif for every observed language. English language is intertwined with complex structured motifs (11, 12 and 13 motif type). Pearsons coefficient correlation is high for English and Italian language for almost every motif type. Croatian language has a low correlation coefficient when compared to other two.

motif, težinski motif, kompleksne mreže, mreže jezika

nije evidentirano