| Carlos Pinheiro |
Network science is the study of connected things. Things can be represented by people, devices, companies,
governments, etc. Connected can be represented by calls, messages, likes, money transferring, references,
geo-positioning, contracts, etc. Networks are constructed from relationships among entities, throughout different types of
events. Network analysis can be used to understand churn, diffuse products, detect fraud, identify anomalies, along with
many other applications. Network optimization searches for optimal solutions to minimize or maximize an objective
function based on a problem described as a graph structure. The objective function can be cost, revenue, time, distance,
and problems are often assigned to supply chain, routing, distribution, etc.
In traditional network analysis, components,
cores, or communities can be used to reveal groups of accounts, persons or organizations that are strongly related. In
network optimization, cycles, cliques, paths, and shortest paths can reveal patterns on the connections that cannot be
highlighted by traditional analysis. Pattern matching can search for specific groups of nodes with specific types of
relations among them. Node similarity searches for entities in the network with similar topology or neighborhood. Network
algorithms can be applied to an endless list of business cases, virtually all industries, to solve problems or to provide
additional insights and features to traditional unsupervised and supervised techniques. |