We propose a methodology for the extraction of local trends from a stream of data. It has been designed to suit the needs of interpretation-oriented visualization and symbolization from ICU monitoring data. After giving implementation details for efficient computation of local trends, we propose the use of a characteristic analysis span for each variable. This characteristic span is obtained from a set of criteria that we compare and evaluate in regard of analysis of ICU monitoring data gathered within the Aiddaig project. The processing results in a rich visual representation and a framework for the local symbolization of the data stream based on its dynamics.
"Trends extracted are linear or quadratic. Another field of application is Intensive Care Unit patient's monitoring, where the aim of a diagnostic system is to recognize on line a change in the patient's health state and to be able to filter false alarms by detecting unphysiological changes (Hunter and McIntosh, 99, Calvelo et al. 2000). Trends extracted in this application field are usually linear. "
[Show abstract][Hide abstract] ABSTRACT: On-line trend extraction is the first step to be achieved by a pattern-matching diagnostic system. Indeed, most pattern-matching diagnostic methods are based on the classification of qualitative or semi-qualitative trends extracted from one or several signals. The relevance of the trend extracted is a key point for the diagnostic system accuracy. This paper presents a trend extraction method which is robust to the presence of artefacts and step-like variations and does not require a priori tuning of the parameters of the method. The parameters are tuned on line by the algorithm itself (auto-tuning method), using a robust estimate of the signal variability. Results obtained on both simulated data and real data show the efficiency of the method.
"Our own research is similar in that we are concerned with the applications of results from machine learning processes to data streams in order to detect adverse clinical condi- tions. Shashar et al.  utilise the work of Calvelo et al.  for decomposing the data stream into trend and stability for each data point. The maximum length of each segment, which gives the best approximation according to the overall piece-wise accuracy and also localised linear accuracy is called the characteristic span. "
[Show abstract][Hide abstract] ABSTRACT: Intelligent clinical data analysis systems require precise qualitative descriptions of data to enable effective and context sensitive interpretation to take place. Temporal abstraction (TA) provides the means to achieve such descriptions, which can then be used as input to a reasoning engine where they are evaluated against a knowledge base to arrive at possible clinical hypotheses. This paper surveys previous research into the development of intelligent clinical data analysis systems that incorporate TA mechanisms and presents research synergies and trends across the research reviewed, especially those associated with the multi-dimensional nature of real-time patient data streams. The motivation for this survey is case study based research into the development of an intelligent real-time, high-frequency patient monitoring system to provide detection of temporal patterns within multiple patient data streams.
The survey was based on factors that are of importance to broaden research into temporal abstraction and on characteristics we believe will assume an increasing level of importance for future clinical IDA systems. These factors were: aspects of the data that is abstracted such as source domain and sample frequency, complexity available within abstracted patterns, dimensionality of the TA and data environment and the knowledge and reasoning underpinning TA processes.
It is evident from the review that for intelligent clinical data analysis systems to progress into the future where clinical environments are becoming increasingly data-intensive, the ability for managing multi-dimensional aspects of data at high observation and sample frequencies must be provided. Also, the detection of complex patterns within patient data requires higher levels of TA than are presently available. The conflicting matters of computational tractability and temporal reasoning within a real-time environment present a non-trivial problem for investigation in regard to these matters. Finally, to be able to fully exploit the value of learning new knowledge from stored clinical data through data mining and enable its application to data abstraction, the fusion of data mining and TA processes becomes a necessity.
Artificial Intelligence in Medicine 02/2007; 39(1):1-24. DOI:10.1016/j.artmed.2006.08.002 · 2.02 Impact Factor
"A patient record is a long, high frequency, multivariate time series data. Classical monitoring improvement techniques are focused on algorithms at a numerical level (Tsien et al., 2000; Calvelo et al., 2000). On the contrary, we propose that symbolic abstraction could bring useful information to data, providing a better understanding of these still poorly formalized data but also, and more deeply, a new approach for monitoring based on dynamic anomaly detection. "
[Show abstract][Hide abstract] ABSTRACT: We work on the design of computerized systems that support experts during their complex and poorly formalized data interpretation process. We consider interpretation as the process by which high level abstracted information is attached to data. We assume that a computer could efficiently helps an expert in such a process via a structural coupling (Maturana and Varela, 1994) based on their interactions. Enaction appears as a stimulating source of inspiration for the design of such systems. Our approach is applied to the exploration of physiological time series acquired from patients in intensive care unit (ICU). Time series interpretation for ICU monitoring systems
Data provided are for informational purposes only. Although carefully collected, accuracy cannot be guaranteed. The impact factor represents a rough estimation of the journal's impact factor and does not reflect the actual current impact factor. Publisher conditions are provided by RoMEO. Differing provisions from the publisher's actual policy or licence agreement may be applicable.