The Use of Variational Bayes Compared to Markov Chain Monte Carlo in The Estimation of Thermal Properties

Estimating the parameters that describe a thermal problem using Bayes statistics requires the specification of appropriate prior probabilities. That is p(P|D) = p(D|P)p(P)/p(D) where P = parameters, D = data and p(P) is the prior probability. For thermal problems this requires prior probabilities for density, specific heat, thermal conductivities, surface convective coefficients, radiative properties, and local heat release, Q. For many problems it is common to choose Gaussian probabilities to represent the errors. If the standard deviation is large, then the predictions can lead to negative values — a result that is not possible except for Q. Variational Bayes (VB) is an alternative to Markov Chain Monte Carlo (MCMC) and assumes that complex distributions p(a,b) can be replaced by factorization, p(a,b) = p(a)p(b), the mean field theory of physics. Overall Variational Bayes is particularly important for posterior probabilities, p(a|D), that have multiple maxima distributions.

Melissococcus plutonius Can Be Effectively and Economically Detected Using Hive Debris and Conventional PCR

European foulbrood (EFB) is an infectious disease of honey bees caused by the bacterium Melissococcus plutonius. A method for DNA isolation and conventional PCR diagnosis was developed using hive debris, which was non-invasively collected on paper sheets placed on the bottom boards of hives. Field trials utilized 23 honey bee colonies with clinically positive symptoms and 21 colonies without symptoms. Bayes statistics were applied to calculate the comparable parameters for EFB diagnostics when using honey, hive debris, or samples of adult bees. The reliability of the conventional PCR was 100% at 6.7 × 103 Colony Forming Unit of M. plutonius in 1 g of debris. The sensitivity of the method for the sampled honey, hive debris, and adult bees was 0.867, 0.714, and 1.000, respectively. The specificity for the tested matrices was 0.842, 0.800, and 0.833. The predictive values for the positive tests from selected populations with 52% prevalence were 0.813, 0.833, and 0.842, and the real accuracies were 0.853, 0.750, and 0.912, for the honey, hive debris, and adult bees, respectively. It was concluded that hive debris can effectively be utilized to non-invasively monitor EFB in honey bee colonies.

SEMG Analysis Basing on AR Model and Bayes Taxonomy

This paper introduced a pattern recognition method based on auto-regression (AR) model and bayes taxonomy. The proposed methodology consists of three steps. In the first step, the paper designs a circuit to collect surface electromyography (SEMG) signal. In the second step, Auto-regressive (AR) modeling in time series has been applied on people’s forearm muscle. So, EMG signal is preprocessed using AR-Model to extract features from MES. After calculated the coefficients of and AR model, we distill the AR coefficients as its eigenvector. In the third step, a bayes statistics algorithm is designed to classify the muscle movement of forearm. This paper finds this method has many advantages such as reducing error recognition rate and has a relative good result. It proves that there are some relations between motion pattern and AR coefficients. At the same time, this paper adopts virtual instrument technology to raise accuracy of measurement, reduce the cost and workload.

An alternative Bayesian statistics for probabilistic earthquake prediction in Mexico, Central and South America

The probabilities of occurrence of strong (M>6.5) earthquakes, in the seismically active regions of Mexico, central and south America, are estimated. The straightforward approach of Bayes statistics is applied in order to search for the inter-arrival times of strong earthquakes in predefined seismic zones of the above referred regions. The method introduced allows to determine the uncertainties involved, which are expressed as percentages of the earthquake mean return period. The determination in this way is very efficient because one may calculate uncertainties on the same time scale. It is also shown that the final maximum Bayesian probabilities of the inter-arrival times in the several seismic zones are dependent on the data set used and particularly on its time length. Comparisons between the predicted and the real time of earthquake occurrences are finally made in order to evaluate the correlation between them.

Calibration and measurement control based on Bayes statistics

The Bayesian methodology described in this paper has the inherent capability of choosing, from calibration-type curves, candidates which are plausible with respect to measured data, expert knowledge and theoretical models (including the nature of the measurement errors). The basic steps of Bayesian calibration are reviewed and possible applications of the results are described in this paper. A calibration related to head-space gas chromatographic data is used as an example of the proposed method. The linear calibration case has been treated with a log-normal distributed measurement error. Such a treatment of noise stresses the importance of modelling the random constituents of any problem.