Variable selection and classification are common objectives in the analysis of high-dimensional data. Most such methods make distributional assumptions that may not be compatible with the diverse families of distributions data can take. A novel Bayesian nonparametric discriminant analysis model that performs both variable selection and classification within a seamless framework is proposed. Pólya tree priors are assigned to the unknown group-conditional distributions to account for their uncertainty, and allow prior beliefs about the distributions to be incorporated simply as hyperparameters. The adoption of collapsed variational Bayes inference in combination with a chain of functional approximations led to an algorithm with low computational cost. The resultant decision rules carry heuristic interpretations and are related to an existing two-sample Bayesian nonparametric hypothesis test. By an application to some simulated and publicly available real datasets, the proposed method exhibits good performance when compared to current state-of-the-art approaches.
An explicit discontinuous deformation analysis (DDA) that uses an explicit time integration procedure and an explicit calculation of interaction forces between blocks is proposed to overcome the limitations of conventional implicit DDA in simulating large-scale problems. The advantages of the explicit DDA are that (1) the global equilibrium equations are unnecessary to be assembled and the solving for unknowns of every block can be performed independently and conveniently, thereby reducing the computational effort and memory requirement; (2) the open-close iteration process is avoided because the interaction forces between blocks are calculated explicitly according to the initial information at the start of the current time step. The efficient parallel computing is very appropriate for the explicit DDA. To further improve its computational efficiency, the explicit DDA is paralleled based on OpenMP. The accuracy of the explicit DDA is verified through several numerical examples with analytical solutions, experimental data or field observation. Further, the computational efficiency is demonstrated by a series of models and the parallel speedup factor on 6 OpenMP threads is approximately 4.2. Conclusively, the explicit DDA is promising for analyzing blocky systems in large scale.
A parallel within the development of flow analysis and the consolidation of Talanta as one of the main journals in analytical chemistry is drawn. Influence of scientific divulgation, meeting organizations, thematic issues devoted to scientific events and Talanta awards in the recent development of flow analysis is emphasized. For didactic purposes, the discussion is focused on three 20-year periods. A scientometric overview demonstrated the consolidation of Talanta as the main journal for divulgation of recent innovations in flow analysis.
Estimating covariance matrices is an important research topic in statistics and finance. A semiparametric model for covariance matrix estimation is proposed. Specifically, the covariance matrix is modeled as a polynomial function of the symmetric adjacency matrix with time varying parameters. The asymptotic properties for the time varying coefficient and the associated semiparametric covariance estimators are established. A Bayesian information criterion to select the order of the polynomial function is also investigated. Simulation studies and an empirical example are presented to illustrate the usefulness of the proposed method.
The objective of this study was to determine the kinetic parameters and apply Markov Chain Monte Carlo (MCMC) simulation to predict the growth of from spores in cooked ground chicken meat during dynamic cooling. Inoculated samples were exposed to various cooling conditions to observe dynamic growth. A combination of 4 cooling profiles was used in one-step inverse analysis with the Baranyi model as the primary model and the cardinal parameters model as the secondary model. Six kinetic parameters of the Baranyi model and the cardinal parameters model, including , , , , , and , were estimated. The estimated , , and were 14.8, 42.9, and 50.5 °C, respectively, with a of 5.25 h and maximum cell density of 8.4 log CFU/g. Correlation analysis showed that both and are weakly correlated to other parameters, while the remaining parameters are mostly mildly to strongly correlated with each other. Although it may be difficult to estimate highly correlated parameters using a single temperature profile, one-step analysis with multiple different temperature profiles helped estimate them successfully. The estimated parameters were used as the prior information to construct the posterior distribution for Bayesian analysis. MCMC simulation was used to predict the bacterial growth using different dynamic temperature profiles for validation of the accuracy of the predictive models. The MCMC simulation results showed that the Bayesian analysis produced more accurate predictions of bacterial growth during cooling than the deterministic method. With Bayesian analysis, the root-mean-square-error (RMSE) of prediction was only 0.1 log CFU/g with all residual errors within ±0.25 log CFU/g. Therefore, Bayesian analysis is recommended for predicting the growth of in cooked meat during cooling.
In this study, a land use suitability analysis was conducted for rural tourism in the Yenice district, located in the north-west of Turkey. As part of the research process involved dividing the area in question into landscape units using GIS and RS techniques. A suitability rating for tourism activities in each landscape unit was obtained by following through the steps of the ELECTRE method, individually repeated for each landscape unit. It is considered that the 1st-, 2nd- and 3rd-degree suitable activities were most relevant in the rating of the nine different tourism activities. Therefore assessments were made on the basis of these first three ranks. As a result of the analysis, from the 1st-degree suitable activities identified, the first three were found to be mountaineering, trekking and wildlife observation. From the 2nd-degree suitable activities, the first three were flora observation, trekking and hiking, and from the 3rd-degree suitable activities, the first three trekking, orienteering and mountaineering.
The efficient estimation of regression coefficients in the longitudinal data analysis requires a correct specification of the covariance structure. Existing approaches usually focus on modeling the mean with specification of certain covariance structures, which may lead to inefficient or biased estimators of parameters in the mean if misspecification occurs. In this article, we propose a novel data-driven approach based on semiparametric varying-coefficient models to model the mean and the covariance simultaneously, motivated by the modified Cholesky decomposition. An iterative estimation method is proposed, consisting of an orthogonality-based technique for parameters, an adaptive jump-preserving estimation method for varying coefficients, a modification of local linear smoothing technique for the autoregressive coefficient function, and a kernel smoothing technique for the variance function. Theoretical properties of the resulting estimators including uniform consistency and asymptotic normality are explicitly studied under certain mild conditions. Simulation studies are carried out to evaluate the efficacy of the proposed methods, and an analysis of a real data example is provided for illustration.
The objective of the current study is to assess the technical performance of Aquifer Thermal Energy Storage (ATES) based on the monitoring data from 73 Dutch ATES systems. With a total abstraction of 30.4 GWh heat and 31.8 GWh cold per year, the average annual amount of supplied thermal energy was measured as 932.8 MWh. The data analysis revealed only small thermal imbalances and small temperature losses during the storage period. The abstraction temperatures are around 10 and 15 C during summer and winter, respectively. However, the temperature difference between the abstraction and injection wells is 3–4 K smaller compared to the optimal design value. This indicates insufficient interaction between the energy system and the subsurface by an inadequate charging of the aquifer. In addition, the amount of stored and abstracted thermal energy is approximately 50% lower than the capacities licensed by the authorities. This results in an unsustainable utilization of the subsurface. Even though ATES technology proved its enormous potential to significantly reduce CO emissions, the operation still can be optimized. This applies in particular to an adequate planning and maintenance of the building energy system and a more efficient use of the available subsurface space.
In electrochemical machining (ECM), the electrolyte flow field has marked influence on the processing stability, efficiency and surface quality. This paper presented a new electrolyte flow mode to improve the flow field distribution in counter-rotating electrochemical machining (CRECM), in which the electrolyte flows into the processing zone from both sides and top of the fixture, and flows out from the bottom of the fixture. The flow field distributions were analyzed and simulated via computational fluid dynamics (CFD) software. The simulation results showed that the low-velocity area in CRECM was eliminated and the velocity distribution was more uniform by using the optimized flow mode. Experiments were also conducted using a specific fixture. The experimental results indicated that a stable machining process with a small inter-electrode gap could be achieved, and the ECM accuracy was remarkably improved. The sidewall taper angles and rounded corners of the machined convex structure were reduced to be only about -0.8° and 0.6 mm.