An algorithmic method for detecting bunching in the value distribution of related variables.
Ioannis KatsarosKostas KaramanisGeorgios KoliasPublished in: SEEDA-CECNSM (2021)
Keyphrases
- computational cost
- synthetic data
- experimental evaluation
- neural network
- similarity measure
- preprocessing
- energy function
- error rate
- segmentation method
- random variables
- detection method
- optimization algorithm
- high precision
- optimization method
- high accuracy
- significant improvement
- prior knowledge
- pairwise
- gaussian distribution
- test data
- density function
- distribution function
- classification method
- clustering method
- support vector machine svm
- theoretical analysis
- edge detection
- probability distribution
- image processing