Large Knowledge Bases and Networks: Results on Analogies and Betweenness Centrality

Betweenness Centrality and Analogy Solver are the main themes of this Thesis Report. In the rst part we study betweenness centrality on general trees with some results regarding the relation between k-betweenness centrality and betweenness centrality. We also study complete binary trees to derive some new formulas with regards to number of shortest paths through a vertex. In the second part we provide an approach which tries to solve analogies using ConceptNet 4.0. A detailed analysis of few examples and results on the data set is also provided.