作者: G. Guimarães , J.-H. Peter , T. Penzel , A. Ultsch
DOI: 10.1016/S0933-3657(01)00089-6
关键词: Knowledge acquisition 、 Abstraction (linguistics) 、 Machine learning 、 Series (mathematics) 、 Knowledge representation and reasoning 、 Computer science 、 Sensitivity (control systems) 、 Grammar induction 、 Artificial neural network 、 Artificial intelligence 、 Identification (information) 、 Natural language processing
摘要: This paper presents a method for the discovery of temporal patterns in multivariate time series and their conversion into linguistic knowledge representation applied to sleep-related breathing disorders. The main idea lies introducing several abstraction levels that allow step-wise identification patterns. Self-organizing neural networks are used discover elementary series. Machine learning (ML) algorithms use results automatically generate rule-based description. At next levels, grammatical rules inferred. covers one ''bottlenecks'' design knowledge-based systems, namely, acquisition problem. An evaluation lead an overall sensitivity 0.762, specificity 0.758.