Any cooperation in multiple-participant decision making (DM) relies on an exchange of individual knowledge pieces and aims. A general methodology of their rational exploitation without calling for an objective mediator is still missing. Desired methodology is proposed for an important particular case, when a participant, performing Bayesian parameter estimation, is offered a model relating the observable data to their past history.The designed solution is based on the so-called fully probabilistic design (FPD) of DM strategies. The result reduces to an 'ordinary' Bayesian estimation if the offered model is the sample probability density function (pdf), i.e. if it provides additional observations.
An exploitation of prior knowledge in parameter estimation becomes vital whenever measured data is not informative enough. Elicitation of quantified prior knowledge is a well-elaborated art in societal and medical applications but not in the engineering ones. Frequently required involvement of a facilitator is mostly unrealistic due to either facilitator's high costs or complexity of modelled relationships that cannot be grasped by humans. This paper provides a facilitator-free approach based on an advanced knowledgesharing methodology. It presents the approach on commonly available types of knowledge and applies the methodology to a normal controlled autoregressive model.
Background: Continuous non-invasive blood pressure (BP) measurement is a desired virtue in clinical practice. Unfortunately, current systems do not allow one for continuous, reliable BP measurement for more than a few hours per day, and they often require a complicated set of sensors to provide the necessary biosignals. Therefore we investigated the possibility of proposing a computational model that would predict the BP from pulse waves recorded in a single spot. Methods: Two experimental circuits were created. One containing a simple plastic tube for model development and a second with a silicone molded patient-specific arterial tree model. The first model served for the measuring of pulse waves under various BP (70–270 mmHg) and heart rate (60–190 beats per minute) values. Four different computational models were used to estimate the BP values from the diastolic time. The most accurate model was further validated using data from the latter experimental circuit containing a molded patient-specific silicone arterial tree. The measured data were averaged over a window of one, three, and five cycles. Two models based on pulse arrival time (PAT) were also analyzed for comparison. Results: The most accurate model exhibits a correlation coefficient of r = 0.967. The Bland–Altman plot revealed standard deviations (SD) between the model predictions and measurement of 10, 8.3, and 7.5 mmHg for the systolic BP and 8.7, 7 and 6.3 mmHg for the diastolic BP (both pressures calculated for the averaging windows of one, three, and five cycles, respectively). The best of the used PAT based model exhibited a SD of 17, 16, and 15 mmHg for the systolic BP and 14, 13, and 12 mmHg for the diastolic BP for the same averaging windows. Discussion: The proposed model showed its capability to predict BP accurately from the shape of the pulse wave measured at a single spot. Its SD was about 50% lower compared to the PAT based models which met the requirements of the Association for the Advancement of Medical Instrumentation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.