Open AccessIdentifying influential variables in complex system: Network topology versus principal component analysis
Author(s) -
Nur Syahidah Yusoff,
Shamshuritawati Sharif
Publication year - 2016
Publication title -
aip conference proceedings
Language(s) - English
Resource type - Conference proceedings
eISSN - 1551-7616
pISSN - 0094-243X
DOI - 10.1063/1.4954628
Subject(s) - principal component analysis , network topology , topology (electrical circuits) , centrality , computer science , component (thermodynamics) , network analysis , reliability (semiconductor) , key (lock) , complex network , covariance , variable (mathematics) , principal (computer security) , data mining , mathematics , artificial intelligence , engineering , statistics , mathematical analysis , power (physics) , physics , computer security , combinatorics , quantum mechanics , world wide web , electrical engineering , thermodynamics , operating system
High dimensional covariance structure can be considered as a complex system that relates each variable to the others in terms of variability. In complex system, identifying influential variables is a very important part of reliability analysis, which has been a key issue in analysing the structural organization of a system. To analyse such complex system, network topology and principal component analysis are constructed to simplify the system. Network topology can be used to simplify the information about the system and centrality measure will be used to interpret the network. In the other hand, the principal component analysis can be used to eliminate the variables that contribute little extra information. An example will be discussed to illustrate the advantage and disadvantage of network topology and principal component analysis and a recommendation will be presented
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom