Assessment of Outliers and Detection of Artifactual Network Segments Using Univariate and Multivariate Dispersion Entropy on Physiological Signals

Network physiology has emerged as a promising paradigm for the extraction of clinically relevant information from physiological signals by moving from univariate to multivariate analysis, allowing for the inspection of interdependencies between organ systems. However, for its successful implementation, the disruptive effects of artifactual outliers, which are a common occurrence in physiological recordings, have to be studied, quantified, and addressed. Within the scope of this study, we utilize Dispersion Entropy (DisEn) to initially quantify the capacity of outlier samples to disrupt the values of univariate and multivariate features extracted with DisEn from physiological network segments consisting of synchronised, electroencephalogram, nasal respiratory, blood pressure, and electrocardiogram signals. The DisEn algorithm is selected due to its efficient computation and good performance in the detection of changes in signals for both univariate and multivariate time-series. The extracted features are then utilised for the training and testing of a logistic regression classifier in univariate and multivariate configurations in an effort to partially automate the detection of artifactual network segments. Our results indicate that outlier samples cause significant disruption in the values of extracted features with multivariate features displaying a certain level of robustness based on the number of signals formulating the network segments from which they are extracted. Furthermore, the deployed classifiers achieve noteworthy performance, where the percentage of correct network segment classification surpasses 95% in a number of experimental setups, with the effectiveness of each configuration being affected by the signal in which outliers are located. Finally, due to the increase in the number of features extracted within the framework of network physiology and the observed impact of artifactual samples in the accuracy of their values, the implementation of algorithmic steps capable of effective feature selection is highlighted as an important area for future research.

Anuran assemblage changes along small-scale phytophysiognomies in natural Brazilian grasslands

We studied the species composition of frogs in two phytophysiognomies (grassland and forest) of a Ramsar site in southern Brazil. We aimed to assess the distribution of species on a small spatial scale and dissimilarities in community composition between grassland and forest habitats. The sampling of individuals was carried out through pitfall traps and active search in the areas around the traps. We evaluated the existence of these differences by using permutational multivariate analysis of variance and multivariate dispersion. We found 13 species belonging to six families. Leptodactylidae and Hylidae were the most representative families. The compositional dissimilarity was higher between the sampling sites from different phytophysiognomies than within the same phytophysiognomy, suggesting that forest and grassland drive anuran species composition differently. Also, the difference in anuran species composition between the sampling sites within the forest was considerably high. Based on our results, we could assume that the phytophysiognomies evaluated here offer quite different colonization opportunities for anurans, especially those related to microhabitat characteristics, such as microclimate variables.

Stocktaking the environmental coverage of a continental ecosystem observation network

Field-based sampling of terrestrial habitats at continental scales is required to build ecosystem observation networks. However, a key challenge for detecting change in ecosystem composition, structure and function is to obtain a representative sample of habitats. Representative sampling contributes to ecological validity when analysing large spatial surveys, but field resources are limited and representativeness may differ markedly from purely practical sampling strategies to statistically rigorous ones. Here, we report a post hoc assessment of the coverage of environmental gradients as a surrogate for ecological coverage by a continental-scale survey of the Australian Terrestrial Ecosystem Research Network (TERN). TERN’s surveillance program maintains a network of ecosystem observation plots that were init ially established in the rangelands through a stratification method (clustering of bioregions by environment) and Ausplots methodology. Subsequent site selection comprised gap filling combined with opportunistic sampling. Firstly, we confirmed that environmental coverage has been a good surrogate for ecological coverage. The cumulative sampling of environments and plant species composition over time were strongly correlated (based on mean multivariate dispersion; r = 0.93). We then compared the environmental sampling of Ausplots to 100,000 background points and a set of retrospective (virtual) sampling schemes: systematic grid, simple random, stratified random, and generalised random-tessellation stratified (GRTS). Differences were assessed according to sampling densities along environmental gradients, and multivariate dispersion (environmental space represented via multi-dimensional scaling). Ausplots outperformed systematic grid, simple random and GRTS in coverage of environmental space (Tukey HSD of mean dispersion, p < .001). GRTS site selection obtained similar coverage to Ausplots when employing the same bioregional stratification. Stratification by climatic zones generated the highest environmental coverage (p < .001), but the resulting sampling densities over-represented mesic coastal habitats. The Ausplots stratification by bioregions implemented under practical constraints represented complex environments well compared to statistically oriented or spatially even samples. However, potential statistical inference and power also depend on spatial and temporal replication, unbiased site selection, and accurate field measurements relative to the magnitude of change. A key conclusion is that environmental, rather than spatial, stratification is required to maximise ecological coverage across continental ecosystem observation networks.

ORGANIZATION OF QUALITY CONTROL AND DIAGNOSTICS PRODUCTION OF STYRENE-ACRYLIC DISPERSION USING ANALYSIS OF VARIANCE

The production of styrene-acrylic dispersion is a complex technological process, which requires constant monitoring. The organization of product quality control has become the main part of the production process and is aimed not only at identifying defects or defective finished products, but also at diagnosing the quality of products in the process of their production, which ensures the advantage of the product in the market and also contributes to its successful marketing. competitive conditions. In various publications, the topic of controlling the process of styrene-acrylic dispersion production has been repeatedly raised. In this paper, a method for organizing the control and diagnostics of the quality of production of a styrene-acrylic dispersion is proposed. For diagnostics, the model of multivariate dispersion analysis was chosen. Based on the visual analysis and the technical characteristics of the product, a rationale is given for the selection of factors that influence quality indicators. The analysis is based on reliable data. The proposed diagnosis has been tested in production.

A Hybrid Vehicle Extraction Approach from Low-Quality LiDAR Data Based on Robust One-Class Support Vector Machine

Vehicle extraction becomes possible as the high-performance airborne light detection and ranging (LiDAR) systems can offer very dense and accurate point cloud, which means the sophisticated objects can be recorded in detail, combined with color information from airborne image, hyperspectral and intensity. However, few studies have investigated in extracting vehicles from LiDAR data only, especially when its quality is low, which is the main difficulty for most LiDAR applications. In this paper, a hybrid approach has been proposed to extract vehicles from low-quality LiDAR data. In order to extract vehicle from low-resolution LiDAR data, a robust one-class support vector machine-minimum covariance determinant (OCSVM-MCD) is proposed based on a multivariate dispersion estimator and weighted strategy. Firstly, the three-dimensional (3D) point dataset is classified into nonterrain and terrain points with progressive morphological filter with a slight improvement. Secondly, nonterrain points are segmented by clustering technique and missing blobs are searched from terrain points. Then, the vehicles are extracted from clustering and searching results by OCSVM-MCD, and a hybrid principle is put forward to improve the extraction result at last. The proposed method has been evaluated with two benchmark datasets from ISPRS, and proved that by the method, most vehicles can be extracted from low-quality LiDAR data with an encouraging result.

A Statistical Test for the Stability of Covariance Structure

The stability of covariance matrix is a major issue in multivariate analysis. As can be seen in the literature, the most popular and widely used tests are Box M-test and Jennrich J-test introduced by Box in 1949 and Jennrich in 1970, respectively. These tests involve determinant of sample covariance matrix as multivariate dispersion measure. Since it is only a scalar representation of a complex structure, it cannot represent the whole structure. On the other hand, they are quite cumbersome to compute when the data sets are of high dimension since they do not only involve the computation of determinant of covariance matrix but also the inversion of a matrix. This motivates us to propose a new statistical test which is computationally more efficient and, if it is used simultaneously with M-test or J-test, we will have a better understanding about the stability of covariance structure. An example will be presented to illustrate its advantage