An Accurate Tax Fraud Classifier with Feature Selection based on Complex Network Node Centrality Measure.
Tales MatosJosé Antônio Fernandes de MacêdoJosé Maria MonteiroFrancesco LettichPublished in: ICEIS (1) (2017)
Keyphrases
- complex networks
- degree centrality
- betweenness centrality
- feature selection
- centrality measures
- social network analysis
- feature set
- network analysis
- community detection
- social networks
- feature space
- support vector machine
- support vector
- network structure
- fraud detection
- community structure
- graph theory
- text categorization
- shortest path
- feature subset
- feature selection and classifier
- scale free
- feature ranking
- feature extraction
- clustering coefficient
- feature selection algorithms
- small world
- classification accuracy
- training set
- biological networks
- link prediction
- wrapper method
- classification models
- machine learning
- information theoretic concepts
- community discovery
- training data
- functional modules
- feature reduction
- information gain
- small world networks
- decision trees
- svm classifier
- selection criterion
- protein interaction networks
- network evolution
- real world networks
- multi class