Pattern Recognition Procedures Research Articles

Design and Implementation of a Fuzzy Classifier for FDI Applied to Industrial Machinery.

In the present work, the design and the implementation of a Fault Detection and Isolation (FDI) system for an industrial machinery is proposed. The case study is represented by a multishaft centrifugal compressor used for the syngas manufacturing. The system has been conceived for the monitoring of the faults which may damage the multishaft centrifugal compressor: instrument single and multiple faults have been considered as well as process faults like fouling of the compressor stages and break of the thrust bearing. A new approach that combines Principal Component Analysis (PCA), Cluster Analysis and Pattern Recognition is developed. A novel procedure based on the statistical test ANOVA (ANalysis Of VAriance) is applied to determine the most suitable number of Principal Components (PCs). A key design issue of the proposed fault isolation scheme is the data Cluster Analysis performed to solve the practical issue of the complexity growth experienced when analyzing process faults, which typically involve many variables. In addition, an automatic online Pattern Recognition procedure for finding the most probable faults is proposed. Clustering procedure and Pattern Recognition are implemented within a Fuzzy Faults Classifier module. Experimental results on real plant data illustrate the validity of the approach. The main benefits produced by the FDI system concern the improvement of the maintenance operations, the enhancement of the reliability and availability of the compressor, the increase in the plant safety while achieving reduction in plant functioning costs.

Toward artificial-intelligence assisted design of experiments

Computer science tools involving differentiable programming today enable the end-to-end modeling of the very complex instruments we develop for fundamental physics. A complete model of an experiment studying fundamental interactions usually involves a surrogate or parametric approximation of data-generating processes, along with models of detector layout, pattern recognition, and inference-extraction procedures. The resulting software pipeline enables the study of the dependence of a performance metric on detector-defining parameters and its maximization by stochastic gradient descent across the very large space of design options. The software may identify new design solutions, as well as indicate optimal compromises between cost and performance.

Classification model with collinear grouping of features

ABSTRACT Pattern recognition procedures operate on data represented as sets of multidimensional feature vectors. A small sample of data appears when the dimension of the feature vectors (number of features) is much larger than the number of feature vectors (objects). Small datasets often emerge in practice, for example in genetics. The design of classification or prognostic models on small data sets requires the development of new types of methods. Methods based on L 1 geometry can play an important role in this regard.

Ecological-Epizootiological Differentiation of Natural Plague Foci

Plague foci differentiation is at the core of surveillance and forecasting of epizootic activity and hence achieving high effectiveness of prophylactic measures for prevention of human plague cases. Improvement of the plague foci differentiation taking into account recent data on their status and functioning is a relevant objective. The aim of the study was to differentiate natural plague foci according to the significance of ecology factors in the dynamics of epizootic activity. Materials and methods. Various models for forecasting epizootic activity of natural plague foci, based on the consistent statistical pattern recognition procedure were applied. 11 plague focal territories for which predictive models had ever been developed were surveyed. Results and discussion. The plague foci differ by impact of biotic and abiotic factors on epizootic activity. If epizootic activity depends on the biotic factors, forecasting can only be made on the basis of the data obtained directly during epizootiological surveillance. If epizootic activity depends on abiotic factors, such plague foci can be more depended on the global climate change. The put forward differentiation of focal areas allows for searching the common patterns of plague epizootic process in foci with different species of hosts and vectors, plague microbe variants and geographic setting, as well as for additional typification alongside the existing ones.

Spatially varying effects of measured confounding variables on disease risk

BackgroundThe presence of considerable spatial variability in incidence intensity suggests that risk factors are unevenly distributed in space and influence the geographical disease incidence distribution and pattern. As most human common diseases that challenge investigators are complex traits and as more factors associated with increased risk are discovered, statistical spatial models are needed that investigate geographical variability in the association between disease incidence and confounding variables and evaluate spatially varying effects on disease risk related to known or suspected risk factors. Information on geography that we focus on is geographical disease clusters of peak incidence and paucity of incidence.MethodsWe proposed and illustrated a statistical spatial model that incorporates information on known or hypothesized risk factors, previously detected geographical disease clusters of peak incidence and paucity of incidence, and their interactions as covariates into the framework of interaction regression models. The spatial scan statistic and the generalized map-based pattern recognition procedure that we recently developed were both considered for geographical disease cluster detection. The Freeman-Tukey transformation was applied to improve normality of distribution and approximately stabilize the variance in the model. We exemplified the proposed method by analyzing data on the spatial occurrence of sudden infant death syndrome (SIDS) with confounding variables of race and gender in North Carolina.ResultsThe analysis revealed the presence of spatial variability in the association between SIDS incidence and race. We differentiated spatial effects of race on SIDS incidence among previously detected geographical disease clusters of peak incidence and incidence paucity and areas outside the geographical disease clusters, determined by the spatial scan statistic and the generalized map-based pattern recognition procedure. Our analysis showed the absence of spatial association between SIDS incidence and gender.ConclusionThe application to the SIDS incidence data demonstrates the ability of our proposed model to estimate spatially varying associations between disease incidence and confounding variables and distinguish spatially related risk factors from spatially constant ones, providing valuable inference for targeted environmental and epidemiological surveillance and management, risk stratification, and thorough etiologic studies of disease.

Settled fast measurement of topological charge by direct extraction of plane wave from vortex beam* *Project supported by the National Natural Science Foundation of China (Grant Nos. 61975015, 61575022, and 61421001) and the Beijing Natural Science Foundation, China (Grant No. L191004).

A method of measuring the vortex beam topological charge (TC) is proposed based on a device that can directly extract the plane wave form from the vortex beam in which the different propagation angles of the plane waves are uniquely related to the different TCs. Then the TC can be obtained by simply comparing the energy values perceived by two fixed sensors in the detection location with the help of twin omnidirectional energy absorbers (OEAs). Because the settled detection relies only on the simple quantitative value at two fixed positions, neither pattern recognition nor field analysis procedure is applied, thus allowing faster measurement. Some features of the methodology are investigated, and the numerical simulations verify the feasibility and robustness of the system.

Handwritten Digit Recognition by using Pattern Recognition & Consensus Clustering

In Big Data, Pattern Recognition and Consensus Clustering procedures have developing significance to the scholastic and expert networks. Today there is an extraordinary worry for ordering the information, as information in wrong classification implies incorrect data, which thus results wastage of resources and hurting the association. Example acknowledgment (PR) helps in maintaining a strategic distance from poor order of information by recognizing the right structure of information in dataset. Perceiving an example is the computerized procedure of finding the specific match and regularities of information, which is firmly identified with Artificial Intelligence and Machine Learning. PR goes about as an essential advance to give bunching since it examinations the structure and vector estimation of every character in dataset. Accord Clustering (CC) additionally called as bunching gatherings, assumes a critical job in arranging and keep up in any sort of information. This is a strategy that joins different bunching answers forget steady, precise and novel outcomes. Right now, actualize PR and CC strategies; we use MNIST dataset which is an enormous database of transcribed digits that is regularly utilized for preparing different frameworks in the field of Machine Learning

Intelligente Störungserkennung einer Werkzeugmaschine

Applications for artificial intelligence in production science are on the brink of industrial implementation. For an anomaly detection in machine tools, fixed threshold limits in signal data are necessary. To develop an autonomous anomaly detection without fixed limits, recurring segments have to be detected. In this article a procedure for online pattern recognition in NC code is described, where the recognized segments will be matched with drive signals. The intelligent system independently learns individual threshold limits. This enables an anomaly detection in online drive signals. The user can classify these faults and receives recommendations for action.

Spatio-temporal traffic queue detection for uninterrupted flows

When traffic flow exceeds capacity because of demand fluctuations, crashes, work zones, and special events, a traffic queue is formed on a highway. Traffic queues cause potentially hazardous situations at the end of the queue where drivers unexpectedly face slowed or stopped traffic while approaching at high speed. Therefore, detecting a queue is vital for protecting it. This study presents a real-time spatio-temporal traffic queue detection algorithm that builds on traffic flow fundamentals combined with a statistical pattern recognition procedure. Using flow-density data, traffic flow phase is classified as either congested or uncongested flow in a probabilistic manner, based on Gaussian mixture models for each location in such a way that detects the traffic phase transitions. The proposed detection algorithm was applied to detect traffic queues using traffic detector data from Interstate 40 in Knoxville, Tennessee. The detection results show that the algorithm detects queues successfully by accounting for varying queueing conditions and different queue types. In addition, the proposed algorithm performed better than speed threshold methods in the literature.

An ensemble procedure for pattern recognition of regional drought

AbstractDrought being a complex natural hazard is challenging to monitor. Although standardized procedures for drought characterization are commonly used, the uncertainty about accurate drought characterization under various procedures always exists. Therefore, for effective drought mitigation policies, it is necessary to compare the performance of new and the existing drought procedures. In this study, we proposed a new procedure—The Drought Intensity Pattern Determinate (DIPD) to examine and compare the spatial‐temporal intensities of various drought episodes based on Standardized Drought Indices (SDI). The DIPD framework provides a new tool—The Drought Concentration Index (DCI) to understand the annual intensity of drought classes by alluding the observed frequencies of drought episodes. At the preliminary level of assessment, DIPD is employed at 52 meteorological stations of Pakistan, and compared three commonly used standardized drought indices: Standardized Precipitation Index (SPI), Standardized Evapotranspiration Index (SPEI) and Standardized Precipitation Temperature Index (SPTI). Exploratory analysis shows a mixed picture as all the drought indices behaving similar for some patterns of drought classes, while some classes show significant discrepancies. In addition, we assessed and compared the spatial distributions using spatial Poisson lognormal model by assuming SDI as a non‐Gaussian process for individual total classes of observed DCI. Results show that the proposed method has ability to incorporate the inner variability in patterns of drought classes. In summary, the DIPD provided a new regional based comparative framework, which accounts the all possible episodes of drought classes and effect of inner discrepancies among quantitative values of SDI.

Combination of vintage and new-fashioned analytical approaches for varietal and geographical traceability of olive oils

There is a great need for having accurate analytical methods able to guarantee the authenticity and traceability of foods, especially for those of high quality and economic value such as extra-virgin olive oil. In the present work, we assessed the potential of combining traditional analytical techniques, based on the characterization of the unsaponifiable fraction, together with novel nuclear magnetic resonance fingerprinting with the aim to investigate the effect of variety and geographical origin on olive oils collected from different locations across the province of Huelva (Spain). Various complementary supervised pattern recognition procedures and machine learning algorithms were then applied to build classification and predictive models. Extra-virgin olive oils were characterized by high concentrations of apparent β-sitosterol (93% of total sterol content), α-tocopherol (representing almost 91% of the total tocopherol fraction), squalene (90% of the total hydrocarbon content), heptacosanol and eicosane (the most abundant aliphatic alcohol and n-alkane, respectively). Furthermore, olive oil classes could be clearly differentiated on the basis of a characteristic chemical pattern, comprising tocopherols, squalene, sterols (campesterol, stigmasterol, β-sitosterol), aliphatic alcohols (heptacosanol, octacosanol) and some nuclear magnetic resonances related to fatty acid chains.

АНАЛИЗ МЕТОДОВ ПОСТКЛАССИФИКАЦИОННОЙ ОБРАБОТКИ МНОГОКАНАЛЬНЫХ ИЗОБРАЖЕНИЙ

The subject matter of the article is the methods of morphological spatial filtering of images in pseudo-colors obtained as a result of statistical segmentation of multichannel satellite images. The aim is to study the effectiveness of various methods of post-classification image processing in order to increase the probability of correct recognition for observed objects. The tasks to be solved are: to select a mathematical model describing the training sets of objects’ classes; to implement the procedure of statistical controlled classification by the maximum likelihood method; to evaluate the results of objects’ recognition on the test image by the criterion of the empirical probability of correct recognition; to formalize the procedures of local object-oriented filtering of a segmented image; to investigate the effectiveness of rank filtering as well as weighted median filtering procedures taking into account the results of the classification by k-nearest neighbors in the filter window. The methods used are methods of empirical distributions’ approximation, statistical recognition methods, methods of probability theory and mathematical statistics, methods of local spatial filtering. The following results were obtained. A method for synthesizing a universal mathematical model has been proposed for describing non-Gaussian signal characteristics of objects on multichannel images based on a multi-dimensional variant of Johnson SB distribution; this model was used for statistical pixel-by-pixel classification of the original satellite image. Algorithms for local post-classification processing in the neighborhood of the selected segments boundaries have been implemented. The analysis of the developed algorithms’ effectiveness based on estimates of classes’ correct recognition probability is performed. Conclusions. The scientific novelty of the results obtained is as follows: combined approaches to the pattern recognition procedures have been further developed – it has been shown that the use of methods of local object-oriented filtering of segmented images allows to reduce the number of point errors for element-wise classification of spatial objects.

Combined machine learning and functional magnetic resonance imaging allows individualized prediction of high-altitude induced psychomotor impairment: The role of neural functionality in putamen and pallidum.

Hypoxia exposure during high-altitude expedition cause psychomotor impairment. Neuroimaging studies indicated that the impairment may be significantly associated with neuron loss and decreased regional homogeneity (ReHo) in several brain regions, suggesting the neural functionality in these regions may be utilized to predict psychomotor impairment under exposure. In this study, 69 subjects come from Shaanxi-Tibet immigrant cohort. Reaction time (RT) tasks were performed to measure the subject's psychomotor function before and after 2-year high-altitude exposure. For each individual, the RT differences between pre-exposure and post-exposure were calculated, which were referred to as "targets" in model establishment. Rs-fMRI data were acquired at the same time with RT tasks. For each individual, the map of ReHo alteration was generated, from which the patterns would be recognized. A pattern recognition procedure was utilized to train and test the predictive models. Two different cross-validation strategies were utilized to evaluate the model performance: leave-one-out cross-validation and four-fold cross-validation. For the models displaying significant R2 and MSE, weight maps were built. As a result, the predictive models were able to decode the changes of simple and recognition reaction time from the alterations of brain activation under the exposure. The regions with highest contributions to the predictions were bilateral putamen and bilateral pallidum, suggesting that predictions were mainly based on the patterns concentrated in these regions. This study was a proof of concept study designed to examine whether individual-level psychomotor impairment under high-altitude exposure could be predicted by a combination of pattern recognition approach and neuroimaging data.

Assessment of muscle fatigue during isometric contraction using autonomic nervous system correlates

Muscle fatigue is a complex phenomenon that results in a reduction of the maximal voluntary force. Measuring muscle fatigue can be a challenging task that may involve the use of intramuscular electrodes (i.e., intramuscular electromyography (EMG)) or complex acquisition techniques. In this study, we propose an alternative non-invasive methodology for muscle fatigue detection relying on the analysis of two autonomic nervous system (ANS) correlates, i.e., the electrodermal activity (EDA) and heart rate variability (HRV) series. Based on standard surface EMG analysis, we divided 32 healthy subjects performing isometric biceps contraction into two groups: a fatigued group and a non-fatigued group.EDA signals were analyzed using the recently proposed cvxEDA model in order to derive phasic and tonic components and extract effective features to study ANS dynamics. Furthermore, HRV series were processed to derive several features defined in the time and frequency domains able to estimate the cardiovascular autonomic regulation. A statistical comparison between the fatigued and the non-fatigued groups was performed for each ANS feature, and two EDA features, i.e., the tonic variability and the phasic response rate, showed significant differences. Moreover, a pattern recognition procedure was applied to the combined EDA-HRV feature-set to automatically discern between fatigued and non-fatigued subjects. The proposed SVM classifier, following a recursive feature elimination stage, exhibited a maximal balanced accuracy of 83.33%. Our results demonstrate that muscle fatigue could be identified in a non-invasive fashion through effective EDA and HRV processing.

Smart energy management algorithm for load smoothing and peak shaving based on load forecasting of an island’s power system

In this study, a novel algorithm for the management of the power flows of an islanded power system was developed, capable of simultaneously achieving steadier conventional unit operation and shaving the demand peak values, for the days of the year that present a night peak in their load curve. The under investigation system is composed of Diesel Generators, a PV farm and a Battery Energy Storage System (BESS) with the power system’s consumption to be relatively higher than its RES production. The proposed algorithm combines the use of a load forecasting methodology, a pattern recognition procedure and a custom optimal power flow scheduling algorithm. The prediction module was based on a feedforward artificial neural network, capable of short-term day ahead load forecasting. The forecasted day ahead load profile was then used as an input to the developed pattern recognition algorithm, in order to be classified based on its load curve shape (pattern). Subsequently, in case that the classification resulted in a clear night peak pattern, it was possible to estimate an hourly based trajectory for the diesel generators operation and derive the BESS charging setpoints, which result in the desired peak shaving and smoothing level simultaneously. In this way, it is possible to replace or substitute the highest power demand with stored renewable energy and to operate the diesel engines as steady as possible, diminishing the ramp up and the steep gradients before the night hours’ peak. The algorithm was integrated in the overall system model in APROS software, where dynamic simulations were performed. The simulation results proved that by applying the proposed algorithm, a combined effect of smoother diesel generator operation and peak shaving with renewable energy is achievable even with the absence of PV overproduction, diminishing the variability of the load to be covered from the conventional units. Such an operation aims at enabling diesel engines to be rated at a lower, than currently, maximum capacity while increasing the share of the renewable energy penetration into the grid.

Pattern versus level: a new look at the personality-entrepreneurship relationship

PurposeBy using a nontraditional configuration approach, the purpose of this paper is to examine the effect of personality on entrepreneurial intention and success. Previous research has focused on why individuals become entrepreneurs and why some are more successful than others. However, most studies have investigated only single factors or primary personality traits. The current study investigates not only the strength of the personality-entrepreneurship link, but also clarifies the nature of the relationship.Design/methodology/approachUsing two independent samples and an innovative regression-based pattern recognition procedure, the study investigates whether the nature of the personality-entrepreneurship relationship is driven by individuals’ absolute trait levels or by their idiosyncratic configuration of personality traits. The non-entrepreneur sample consisted of 225 business students in Eastern China, specializing in a variety of business subjects. The entrepreneur sample consisted of 120 business owners in a university entrepreneurial park in Eastern China.FindingsResults support hypotheses that the two different types of entrepreneurship criteria are predicted by different personality profile effects. Entrepreneurial intentions are driven by individuals’ personality patterns (peaks and valleys in profiles). In contrast, entrepreneurial success is driven by personality levels (individuals’ relative standing on personality traits compared to other entrepreneurs).Research limitations/implicationsThe findings enrich the understanding of entrepreneurial personality. The more significant contribution of the present study was that it differentiated between personality profile pattern and level effects and investigated whether the nature of the personality-entrepreneurship relationship is driven by individuals’ absolute trait levels (i.e. how high or low they score compared to others), or by their idiosyncratic configuration of personality traits (i.e. their strengths and weaknesses).Practical implicationsThe findings of this study may help entrepreneurs to figure out how to be successful running their own businesses, if they are not graced with a personality pattern that is not favorable to entrepreneurship. In addition, these findings can help entrepreneurship educators to understand how best to train entrepreneurs.Originality/valueThe adoption of both person-centered approach and process perspective of entrepreneurship allowed this study to make major contributions to entrepreneurial personality research.

Data-driven monitoring of the gearbox using multifractal analysis and machine learning methods

Data-driven diagnostic methods allow to obtain a statistical model of time series and to identify deviations of recorded data from the pattern of the monitored system. Statistical analysis of time series of mechanical vibrations creates a new quality in the monitoring of rotating machines. Most real vibration signals exhibit nonlinear properties well described by scaling exponents. Multifractal analysis, which relies mainly on assessing local singularity exponents, has become a popular tool for statistical analysis of empirical data. There are many methods to study time series in terms of their fractality. Comparing computational complexity, a wavelet leaders algorithm was chosen. Using Wavelet Leaders Multifractal Formalism, multifractal parameters were estimated, taking them as diagnostic features in the pattern recognition procedure, using machine learning methods. The classification was performed using neural network, k-nearest neighbours’ algorithm and support vector machine. The article presents the results of vibration acceleration tests in a demonstration transmission system that allows simulations of assembly errors and teeth wear.

Горизонтальне та вертикальне масштабування методів машинного навчання

The main stages of Machine Learning Pipelines are considered in the paper, such as: train data collection and storage, training and scoring. The effect of the Big Data phenomenon on each of the stages is discussed. Different approaches to efficient organization of computation are on each of the stage are evaluated. In the first part of the paper we introduce the notion of horizontal and vertical scalability together with corresponding cons and pros. We consider some limitations of scaling, such as Amdahl's law. In the second part of the paper we consider scalability of data storage routines. First we discuss relational databases and scalability limitations related to ACID guarantees, which such database satisfy. Then we consider horizontally scalable non-relational databases, so called NoSQL databases. We formulate CAP-theorem as a fundamental limitation of horizontally scalable databases. The third part of the paper is dedicated to scalability of computation based on the MapReduce programming model. We discuss some implementations of this programming model, such as Hadoop and Spark together with some basic principles which they are based on. In the fourth part of the article we consider various approaches towards scaling of Machine Learning methods. We give the general statement of Machine Learning problem. Then we show how MapReduce programming model can be applied for horizontal scaling of Machine Learning methods on the example of Bayessian pattern recognition procedure. On the example of Deep Neural Networks we discuss Machine Learning methods which are not horizontally scalable. Then we consider some approaches towards vertical scaling of such methods based on GPU’s and the TensorFlow programming model.

Forecasting of epizootic Activity of the Central Caucasian natural High-Mountain Plague Focus

Central-Caucasian natural plague focus was permanently epizootically active since its discovering in 1971 till 2007. Inter-epizootic period has been in progress since 2008. It was not possible to isolate agent strains from field material. Therefore a forecast for focus activation is a relevant task, especially against the background of registered plague cases in humans in 2014–2016. Objective of the study was to create a forecasting model for quantitative prediction of possible activation or maintenance of inter-epizootic period. Materials and methods. We used archival data of Kabardino-Balkar Plague Control Station: journals of rodents’ autopsy, annual reports on epizootiological surveillance, meteorological data from meteostation “Kislovodsk” over the period of 1989–2017, and our epidemiological data for the period 2010 to 2017. We applied Spearman nonparametric correlation analysis, regression analysis, including principal component method, quarterly analysis, and inhomogeneous sequential pattern recognition procedures for statistical processing. Results and discussion. We have designed statistical model which provides for forecasting of plague focus epizootic activity proactively, a year in advance and 99 % probability or higher. The model was tested on retrospective data over the course of 7 years. All predictions were correct. The operational forecasts from 2015 to 2017 proved right too. However there is a possibility of fast changes in the ecology system conditions of the Central-Caucasian natural plague focus because of the global warming. Thereby the forecasting model will be annually checked for informative value of the predictors and, if necessary, adjusted accordingly.

Enhancing Residential Water End Use Pattern Recognition Accuracy Using Self-Organizing Maps and K-Means Clustering Techniques: Autoflow v3.1

The aim of residential water end-use studies is to disaggregate water consumption into different water end-use categories (i.e., shower, toilet, etc.). The authors previously developed a beta application software (i.e., Autoflow v2.1) that provides an intelligent platform to autonomously categorize residential water consumption data and generate management analysis reports. However, the Autoflow v2.1 software water end use event recognition accuracy achieved was between 75 to 90%, which leaves room for improvement. In the present study, a new module augmented to the existing procedure improved flow disaggregation accuracy, which resulted in Autoflow v3.1. The new module applied self-organizing maps (SOM) and K-means clustering algorithms for undertaking an initial pre-grouping of water end-use events before the existing pattern recognition procedures were applied (i.e., ANN, HMM, etc.) For validation, a dataset consisting of over 100,000 events from 252 homes in Australia were employed to verify accuracy improvements derived from augmenting the new hybrid SOM and K-means algorithm techniques into the existing Autoflow v2.1 software. The water end use event categorization accuracy ranged from 86 to 94.2% for the enhanced model (Autoflow v3.1), which was a 1.7 to 9% improvement on event categorization.