Loading...

Noise-tolerant model selection and parameter estimation for complex networks

Aliakbary, S ; Sharif University of Technology | 2015

943 Viewed
  1. Type of Document: Article
  2. DOI: 10.1016/j.physa.2015.02.032
  3. Publisher: Elsevier , 2015
  4. Abstract:
  5. Real networks often exhibit nontrivial topological features that do not occur in random graphs. The need for synthesizing realistic networks has resulted in development of various network models. In this paper, we address the problem of selecting and calibrating the model that best fits a given target network. The existing model fitting approaches mostly suffer from sensitivity to network perturbations, lack of the parameter estimation component, dependency on the size of the networks, and low accuracy. To overcome these limitations, we considered a broad range of network features and employed machine learning techniques such as genetic algorithms, distance metric learning, nearest neighbor classification, and artificial neural networks. Our proposed method, which is named ModelFit, outperforms the state-of-the-art baselines with respect to accuracy and noise tolerance in different network datasets
  6. Keywords:
  7. Model selection ; Parameter estimation ; Social networks ; Artificial intelligence ; Genetic algorithms ; Learning systems ; Neural networks ; Parameter estimation ; Social networking (online) ; Topology ; Distance Metric Learning ; Machine learning techniques ; Model Selection ; Model-fitting approach ; Nearest neighbor classification ; Network models ; State of the art ; Topological features ; Complex networks
  8. Source: Physica A: Statistical Mechanics and its Applications ; Volume 427 , 2015 , Pages 100-112 ; 03784371 (ISSN)
  9. URL: http://www.sciencedirect.com/science/article/pii/S0378437115001363