Statistical Pattern Recognition Methods Research Articles

A novel portable smart phone sensing platform based on a supramolecular fluorescence probe for quick visual quantitative detection of picric acid.

Picric acid (PA) is a lethal explosive substance that is easily soluble in water and harmful to the environment. Here, a supramolecular polymer material BTPY@Q[8] with aggregation induced emission (AIE) was prepared by supramolecular self-assembly of cucurbit uril (Q[8]) and 1,3,5-tris[4-(pyridin-4-yl) phenyl] benzene derivative (BTPY), which exhibited aggregation-induced fluorescence enhancement. To this supramolecular self-assembly, the addition of a number of nitrophenols was found to have no obvious effect on the fluorescence, however on addition of PA, the fluorescence intensity underwent a dramatic quench. For PA, BTPY@Q[8] had sensitive specificity and effective selectivity. Based on this, a quick and simple on-site visual PA fluorescence quantitative detection platform was developed using smart phones, and the platform was used to monitor temperature. Machine learning (ML) is a popular pattern recognition technology, which can accurately predict the results from data. Therefore, ML has much more potential for analyzing and improving sensing data than the widely used statistical pattern recognition method. In the field of analytical science, the sensing platform offers a reliable method for the quantitative detection of PA that can be applied to other analytes or micropollutant screening.

Quantitative EEG for early differential diagnosis of dementia with Lewy bodies.

Differentiating between the two most common forms of dementia, Alzheimer's dementia and dementia with Lewy bodies (DLB) remains difficult and requires the use of invasive, expensive, and resource-intensive techniques. We aimed to investigate the sensitivity and specificity of electroencephalography quantified using the statistical pattern recognition method (qEEG-SPR) for identifying dementia and DLB. Thirty-two outpatients and 16 controls underwent clinical assessment (by two blinded neurologists), EEG recording, and a 6-month follow-up clinical assessment. EEG data were processed using a qEEG-SPR protocol to derive a Dementia Index (positive or negative) and DLB index (positive or negative) for each participant which was compared against the diagnosis given at clinical assessment. Confusion matrices were used to calculate sensitivity, specificity, and predictive values for identifying dementia and DLB specifically. Clinical assessment identified 30 cases of dementia, 2 of which were diagnosed clinically with possible DLB, 14 with probable DLB and DLB was excluded in 14 patients. qEEG-SPR confirmed the dementia diagnosis in 26 out of the 32 patients and led to 6.3% of false positives (FP) and 9.4% of false negatives (FN). qEEG-SPR was used to provide a DLB diagnosis among patients who received a positive or inconclusive result of Dementia index and led to 13.6% of FP and 13.6% of FN. Confusion matrices indicated a sensitivity of 80%, a specificity of 89%, a positive predictive value of 92%, a negative predictive value of 72%, and an accuracy of 83% to diagnose dementia. The DLB index showed a sensitivity of 60%, a specificity of 90%, a positive predictive value of 75%, a negative predictive value of 81%, and an accuracy of 75%. Neuropsychological scores did not differ significantly between DLB and non- DLB patients. Head trauma or story of stroke were identified as possible causes of FP results for DLB diagnosis. qEEG-SPR is a sensitive and specific tool for diagnosing dementia and differentiating DLB from other forms of dementia in the initial state. This non-invasive, low-cost, and environmentally friendly method is a promising diagnostic tool for dementia diagnosis which could be implemented in local care settings.

Structural health monitoring for corrosion induced thickness loss in marine plates subjected to random loads

In the highly corrosive environment where marine structures operate, the current industry practice is to overcompensate for corrosion induced thickness loss (CITL) and replace any parts that score below specified allowable limits during scheduled maintenance. Hence, there would be immediate benefits from the implementation of a Structural Health Monitoring (SHM) system that would allow for predictive, condition-based maintenance. The objective of this study is to investigate and assess the effectiveness of state-of-the-art statistical pattern recognition (SPR) and machine learning (ML) methods in association with alternative sensor grid architectures as an SHM scheme for detecting CITL, under highly variable operational conditions. For this purpose, a simple rectangular plate at different corrosion levels (uniform and pitting) was considered as a reference structural element, which was subjected to a stochastic pressure load. Strain response data were produced using Finite Element (FE) simulations and were treated under a probabilistic framework. Elements from detection theory and ML were taken under consideration in order to construct alternative detectors and assess their performance.

Applications of Electronic Nose Coupled with Statistical and Intelligent Pattern Recognition Techniques for Monitoring Tea Quality: A Review

Tea is the most widely consumed non-alcoholic beverage worldwide. In the tea sector, the high demand for tea has led to an increase in the adulteration of superior tea grades. The procedure of evaluating tea quality is difficult to assure the highest degree of tea safety in the context of consumer preferences. In recent years, the advancement in sensor technology has replaced the human olfaction system with an artificial olfaction system, i.e., electronic noses (E-noses) for quality control of teas to differentiate the distinct aromas. Therefore, in this review, the potential applications of E-nose as a monitoring device for different teas have been investigated. The instrumentation, working principles, and different gas sensor types employed for E-nose applications have been introduced. The widely used statistical and intelligent pattern recognition methods, namely, PCA, LDA, PLS-DA, KNN, ANN, CNN, SVM, etc., have been discussed in detail. The challenges and the future trends for E-nose devices have also been highlighted. Overall, this review provides the insight that E-nose combined with an appropriate pattern recognition method is a powerful non-destructive tool for monitoring tea quality. In future, E-noses will undoubtedly reduce their shortcomings with improved detection accuracy and consistency by employing food quality testing.

An MPPCA-based approach for anomaly detection of structures under multiple operational conditions and missing data

Structural anomaly detection based on the structural health monitoring (SHM) data has attracted significant attention owing to its important role in the early warning of structural damage to existing civil structures. Data-driven approaches, where damage-sensitive features are extracted directly from the SHM data using statistical pattern recognition (SPR) techniques without physical models of structures, have been widely studied. Principal component analysis (PCA) and probabilistic PCA (PPCA) are powerful and efficient SPR methods for linear or weakly nonlinear cases. However, some special structures may be subjected to multiple operational conditions, wherein structural configurations such as geometry and mass distribution may change due to the movement of parts or the whole structure, as in retractable roof structures. These changes may give erroneous results in the SPR of the SHM data and eventually in the anomaly detection by a single PCA or PPCA model. This paper presents an improved approach using a mixture of probabilistic principal component analysis (MPPCA) for the anomaly detection of structures under multiple operational conditions with missing measurement data. First, the baseline MPPCA model was constructed for stress data collected under healthy conditions, where the estimation of the MPPCA parameters was reformulated for the missing data cases. Second, three anomaly statistics were presented for newly monitored incomplete data to detect and localize structural anomalies. The probability distributions of the anomaly statistics were estimated to obtain thresholds for outlier detection. Finally, the effectiveness of the MPPCA-based method was investigated by applying the method to the anomaly detection of a retractable roof structure with numerically simulated and real monitored stress data.

An investigation of rail failure due to wear using statistical pattern recognition techniques

The purpose of controlling and detecting errors in transportation systems is to ensure the correct operation of the system. The purpose of this paper is to identify wear in rails by using statistical pattern recognition methods. Due to the low accuracy of some of the current methods designing and introducing an easy way to be used in this area is required. The data of this paper were obtained by the field observation and using image captions of without worn profile and worn profile of railway rails, then the wear phenomenon was identified using the image processing method. Comparing the results of wear identification using pattern recognition techniques is in accordance with field observations. The true positive rate (sensitivity) is calculated and the accuracy of the proposed method indicates the validity of the proposed model.

Data-driven online detection of tip wear in tip-based nanomachining using incremental adaptive support vector machine

This paper presented an online statistical pattern recognition method to detect the severe tip wear in nanomachining using an incremental adaptive support vector machine (IASVM). From the time series data of the collected nanomachining force, the two feature variables (i.e., peak-to-peak force and variance of each machining cycle) were calculated to classify the state of the nanomachining process. To enable online detection, the data was collected in the form of a moving window, sliding once every 0.2 s to add 20 sets of new feature data to update the data set each time. For each data window, the support vector machine (SVM) was restructured by modifying the regularization parameter and the Kernel parameter adaptively. The solution structure of the updated SVM was calculated to classify the incremental data and the disturbing data with the best accuracy. The number of tip failure points in each window was counted to determine the tip wear severity and the moment when the AFM tip needs to be replaced by monitoring the trend of tip failure points. From experimental data, it was shown that the average accuracy of IASVM's recognition of the tip wear conditions was 95% and the average calculation time was 0.166 s, which makes it a promising approach for online detection of tip damage in nanomachining process.

Anti-interference recognition method of aerial infrared targets based on the Bayesian network

The development of infrared decoy countermeasure technologies has resulted in the complexity of the air combat environment. Therefore, higher requirements are put forward considering the infrared imaging-guided air-to-air missile anti-interference target recognition technology. The release of infrared decoys destroys the completeness, saliency and stability of target features. It is difficult for statistical pattern recognition methods based on feature fusion and matching to cover all of the confrontation conditions. In the present study, a Bayesian probabilistic recognition model is proposed that deals with uncertain information about targets and decoys. The proposed model deals with the interference of the air battlefield with the confrontation environment. Moreover, it simulates some functions of the human visual cognition and improves the target recognition ability when interference exists. Furthermore, numerous simulation data samples are utilized to solve the network structure as well as parameters of the Bayesian recognition model and the probabilistic definitions of the target and interference are achieved. Finally, a new aerial infrared target recognition algorithm is constructed based on prior information and the probability recognition model. Results of simulation experiments indicate that the recognition rate of the anti-interference recognition algorithm based on the Bayesian network reaches 90.73%, which is 6.905% higher than that of the anti-interference recognition algorithm based on Naive Bayes classifier in the tested air combat anti-interference simulation image dataset. Moreover, it can solve the problem of anti-interference target recognition to a certain extent such as false targets and target occlusion.

Early damage assessment in large-scale structures by innovative statistical pattern recognition methods based on time series modeling and novelty detection

Time series analysis and novelty detection are effective and promising methods for data-driven structural health monitoring (SHM) based on the statistical pattern recognition paradigm. However, processing substantially large volumes of vibration measurements may represent a serious limitation, especially for long-term SHM programs of large-scale civil structures. Moreover, shortcomings like the choice of an appropriate time series model in an automatic manner, the determination of optimal orders of the identified model and the classification of random high-dimensional features for damage detection, can strongly affect the performance of these approaches. This study is intended to propose statistical pattern recognition methods regarding time series modeling for feature extraction and novelty detection in feature classification in the presence of big data. These methods include an automatic model identification algorithm, an improved order determination approach and a hybrid distance-based novelty detection through a combination of Partition-based Kullback-Leibler divergence and Mahalanobis-squared distance. Experimental datasets relevant to a cable-stayed bridge are considered to validate the effectiveness of the proposed methods. Results demonstrate that: the AutoRegressive-AutoRegressive with eXogenous input (AR-ARX) model turns out to be the most suitable representation for feature extraction; the orders of this model are efficiently and automatically determined; the proposed novelty detection approach is highly successful in detecting damage, even in case of large volumes of random high-dimensional features.

The Power of EEG to Predict Conversion from Mild Cognitive Impairment and Subjective Cognitive Decline to Dementia

Introduction: The aim of this study was to examine if quantitative electroencephalography (qEEG) using the statistical pattern recognition (SPR) method could predict conversion to dementia in patients with subjective cognitive decline (SCD) and mild cognitive impairment (MCI). Methods: From 5 Nordic memory clinics, we included 47 SCD patients, 99 MCI patients, and 67 healthy controls. EEGs analyzed with the SPR method together with clinical data recorded at baseline were evaluated. The patients were followed up for a mean of 62.5 (SD 17.6) months and reexamined. Results: Of 200 participants with valid clinical information, 70 had converted to dementia, and 52 had developed Alzheimer’s disease. Receiver-operating characteristic analysis of the EEG results as defined by a dementia index (DI) ranging from 0 to 100 revealed that the area under the curve was 0.78 (95% CI 0.70–0.85), corresponding to a sensitivity of 71%, specificity of 69%, and accuracy of 69%. A logistic regression analysis showed that by adding results of a cognitive test at baseline to the EEG DI, accuracy could improve. Conclusion: We conclude that applying qEEG using the automated SPR method can be helpful in identifying patients with SCD and MCI that have a high risk of converting to dementia over a 5-year period. As the discriminant power of the method is of moderate degree, it should be used in addition to routine diagnostic methods.

Structural damage identification via physics-guided machine learning: a methodology integrating pattern recognition with finite element model updating

Structural health monitoring methods are broadly classified into two categories: data-driven methods via statistical pattern recognition and physics-based methods through finite elementmodel updating. Data-driven structural health monitoring faces the challenge of data insufficiency that renders the learned model limited in identifying damage scenarios that are not contained in the training data. Model-based methods are susceptible to modeling error due to model idealizations and simplifications that make the finite element model updating results deviate from the truth. This study attempts to combine the merits of data-driven and physics-based structural health monitoring methods via physics-guided machine learning, expecting that the damage identification performance can be improved. Physics-guided machine learning uses observed feature data with correct labels as well as the physical model output of unlabeled instances. In this study, physics-guided machine learning is realized with a physics-guided neural network. The original modal-property based features are extended with the damage identification result of finite element model updating. A physics-based loss function is designed to evaluate the discrepancy between the neural network model output and that of finite element model updating. With the guidance from the scientific knowledge contained in finite element model updating, the learned neural network model has the potential to improve the generality and scientific consistency of the damage detection results. The proposed methodology is validated by a numerical case study on a steel pedestrian bridge model and an experimental study on a three-story building model.

ELECTRO-MECHANICAL IMPEDANCE BASED SHM OF THICK STRUCTURES IN BROAD-BAND FREQUENCY SPECTRUM

In this study, theoretical and experimental studies were conducted in terms of electro-mechanical impedance spectroscopy (EMIS) to monitor the structural health status of thick metallic bar structures in broad-band frequency in the range up to 15 MHz. Thus, in-plane (longitudinal-wise) and out-of-plane (thickness-wise) mode EMI spectra were included in this range of frequency. Then, the thickness mode EMIS peaks were focused for pristine thick steel bar as a baseline and for bars with simulated damage in different depths and different locations. For theoretical part of the study, coupled field finite element analyses (CF-FEA) were carried out by modeling piezoelectric wafer active sensor (PWAS) bonded on steel bars in pristine and in damaged status. The thickness mode EMIS results from both CF-FEA simulations and EMIS measurements were analyzed to investigate the effects of the damage location and size variations through a statistical pattern recognition method as damage index such as the root mean square deviation, RMSD, method. It was found that RMSD values were similar trend for experimental and CF-FEA.

Ultrafast detection of infectious bacteria using optoelectronic nose based on metallic nanoparticles

Developing an accurate and sensitive method for detection of waterborne pathogens and urinary tract infection are important environmental and clinical challenges. We report a colorimetric assay based on arrays of nanoparticles for identification of 10 infectious bacteria strains belonging to both gram positive and gram-negative species. The detection system works based on the interaction of the volatile metabolites emitted from bacteria with the arrays of metallic nanoparticles deposited on paper substrate. The color changes, monitored by a flatbed scanner, are analyzed by statistical pattern recognition methods. A unique colorimetric pattern is achieved for each bacteria strain. We could detect bacteria at very low concentrations (around 100 bacteria/mL) in less than 10 min, which is the fastest method reported ever. The sensor can be used successfully for detection of bacteria in drinking water as well as human urine (containing 177 healthy and 123 patient samples) in less than 50 min, which is dramatically faster than urine culture. The provided detection system opens a new way to develop simple, small, sensitive and rapid devices for detecting pathogens in various clinical samples.

Genetic algorithm-based parameter optimization for EO-1 Hyperion remote sensing image classification

ABSTRACT Because hyperspectral remote sensing images have high band dimension and limited training samples, it is hard to achieve acceptable classification results using conventional statistical pattern recognition methods. Nevertheless, support vector machine (SVM) is suitable for hyperspectral image classification owing to its good generalization ability and minimal structural risk. Aimed to address the problems of large subjectivity, excessive time consumption, and low optimization accuracy in traditional SVM parameter selection, we applied genetic algorithm to optimize the model parameters. Using EO-1 Hyperion data as an example, an SVM object classification algorithm was proposed based on genetic algorithm optimization. To determine the effectiveness and superiority of this algorithm, a comprehensive evaluation and comparative analysis were performed with the classification effects of the cross-validation and maximum likelihood methods, respectively. Experimental results demonstrated that the proposed algorithm can effectively avert the aimlessness of artificial parameter selection and automatically optimize the SVM model parameters. It achieved relatively high overall classification accuracy (91.23%), which was 16.42% and 4.48% higher than the maximum likelihood method and cross-validation method, respectively.

Machine Learning-Assisted Array-Based Biomolecular Sensing Using Surface-Functionalized Carbon Dots.

Fluorescent array-based sensing is an emerging differential sensing platform for sensitive detection of analytes in a complex environment without involving a conventional "lock and key" type-specific interaction. These sensing techniques mainly rely on different optical pattern generation from a sensor array and their pattern recognition to differentiate analytes. Currently emerging, compelling pattern-recognition method, Machine Learning (ML), enables a machine to "learn" a pattern by training without having the recognition method explicitly programmed into it. Thus, ML has an enormous potential to analyze these sensing data better than widely used statistical pattern-recognition methods. Here, an array-based sensor using easy-to-synthesize carbon dots with varied surface functionality is reported, which can differentiate between eight different proteins at 100 nM concentration. The utility of using machine learning algorithms in pattern recognition of fluorescence signals from the array has also been demonstrated. In analyzing the array-based sensing data, Machine Learning algorithms like "Gradient-Boosted Trees" have achieved a 100% prediction efficiency compared to inferior-performing classical statistical method "Linear Discriminant Analysis".

An Improved Particle Filter With a Novel Hybrid Proposal Distribution for Quantitative Analysis of Gold Immunochromatographic Strips

In this paper, a novel statistical pattern recognition method is proposed for accurately segmenting test and control lines from the gold immunochromatographic strip (GICS) images for the benefits of quantitative analysis. A new dynamic state-space model is established, based on which the segmentation task of test and control lines is transformed into a state estimation problem. Especially, the transition equation is utilized to describe the relationship between contour points on the upper and the lower boundaries of test and control lines, and a new observation equation is developed by combining the contrast of between-class variance and the uniformity measure. Then, an innovative particle filter (PF) with a hybrid proposal distribution, namely, deep-belief-network-based particle filter (DBN-PF) is put forward, where the deep belief network (DBN) provides an initial recognition result in the hybrid proposal distribution, and the particle swarm optimization algorithm moves particles to regions of high likelihood. The performance of proposed DBN-PF method is comprehensively evaluated on not only an artificial dataset but also the GICS images in terms of several indices as compared to the PF and DBN methods. It is demonstrated via experiment results that the proposed approach is effective in quantitative analysis of GICS.

Quantitative Electroencephalography Analyzed by Statistical Pattern Recognition as a Diagnostic and Prognostic Tool in Mild Cognitive Impairment: Results from a Nordic Multicenter Cohort Study

Aim: To examine diagnostic and prognostic potential of quantitative electroencephalography (qEEG) analyzed by the statistical pattern recognition (SPR) method in patients with cognitive impairment. We compared the differential diagnostic ability of SPR to visual EEG analysis. Correlation between SPR findings and cerebrospinal fluid (CSF) Alzheimer disease (AD) biomarkers were evaluated. Methods: It is a multicenter cohort study involving 129 patients, (mild cognitive impairment [MCI], AD, and healthy controls). Standardized EEG was performed at baseline. Patients were continuously clinically evaluated. Results: Receiver Operating Characteristic curves showed a low discriminative ability of SPR and no ability to predict clinical progression in patients with MCI. Moderate correlation between SPR analysis and CSF AD biomarkers was found. Conclusion: The diagnostic and prognostic abilities of qEEG were low. The SPR method was superior to the visual EEG analysis. The qEEG method correlates well to CSF AD biomarkers, suggesting association with pathology in AD.

Data-driven damage diagnosis under environmental and operational variability by novel statistical pattern recognition methods

Feature extraction by time-series analysis and decision making through distance-based methods are powerful and efficient statistical pattern recognition techniques for data-driven structural health monitoring. The motivation of this article is to propose an innovative residual-based feature extraction approach based on AutoRegressive modeling and a novel statistical distance method named as Partition-based Kullback–Leibler Divergence for damage detection and localization by using randomly high-dimensional damage-sensitive features under environmental and operational variability. The key novel element of the proposed feature extraction approach is to establish a two-stage offline and online learning algorithms for extracting the residuals of AutoRegressive model as the main damage-sensitive features. This technique brings the great benefit of reducing the computational time and storage space for feature extraction in long-term monitoring conditions. The major contribution of Partition-based Kullback–Leibler Divergence method is to exploit a partitioning strategy for dividing random features into individual partitions and utilize numerical information of partitioning in distance calculation rather than directly applying random samples. Dealing with the major challenging issue of using the high-dimensional features in decision making and applicability to both correlated and uncorrelated random datasets are the main advantages of Partition-based Kullback–Leibler Divergence method. The accuracy and reliability of the proposed approaches are experimentally validated by two well-known benchmark structures. The stationarity and linearity of measured vibration responses for using in AutoRegressive modeling are evaluated by two hypothesis tests. Comparative studies are also conducted to demonstrate the superiority of the proposed methods over some exciting state-of-the-art techniques. Results show that the methods presented here succeed in detecting and locating damage and make time-saving and efficient tools for feature extraction and damage diagnosis.

Using a Hidden Markov Model to Analyze the Flood-Season Rainfall Pattern and Its Temporal Variation over East China

The homogeneous hidden Markov model (HMM), a statistical pattern recognition method, is introduced in this paper. Based on the HMM, a 53-yr record of daily precipitation during the flood season (April–September) at 389 stations in East China during 1961–2013 is classified into six patterns: the South China (SC) pattern, the southern Yangtze River (SY) pattern, the Yangtze–Huai River (YH) pattern, the North China (NC) pattern, the overall wetter (OW) pattern, and the overall drier (OD) pattern. Features of the transition probability matrix of the first four patterns reveal that 1) the NC pattern is the most persistent, followed by the YH, and the SY is the least one; and 2) there exists a SY–SC–SY–YH–NC propagation process for the rain belt over East China during the flood season. The intraseasonal variability in the occurrence frequency of each pattern determines its start and end time. Furthermore, analysis of interdecadal variability in the occurrence frequency of each pattern in recent six decades has identified three obvious interdecadal variations for the SC, YH, and NC patterns in the mid–late 1970s, the early 1990s, and the late 1990s. After 2000, the patterns concentrated in the southern region play a dominant role, and thus there maintains a “flooding in the south and drought in the north” rainfall distribution in eastern China. In summary, the HMM provides a unique approach for us to obtain both spatial distribution and temporal variation features of flood-season rainfall.

Statistical‐based approach for driving style recognition using Bayesian probability with kernel density estimation

Driving styles have a great influence on vehicle fuel economy, active safety, and drivability. To recognize driving styles of path-tracking behaviors for different divers, a statistical pattern-recognition method is developed to deal with the uncertainty of driving styles or characteristics based on probability density estimation. First, to describe driver path-tracking styles, vehicle speed and throttle opening are selected as the discriminative parameters, and a conditional kernel density function of vehicle speed and throttle opening is built, respectively, to describe the uncertainty and probability of two representative driving styles, e.g., aggressive and normal. Meanwhile, a posterior probability of each element in feature vector is obtained using full Bayesian theory. Second, a Euclidean distance method is involved to decide to which class the driver should be subject instead of calculating the complex covariance between every two elements of feature vectors. By comparing the Euclidean distance between every elements in feature vector, driving styles are classified into seven levels ranging from low normal to high aggressive. Subsequently, to show benefits of the proposed pattern-recognition method, a cross-validated method is used, compared with a fuzzy logic-based pattern-recognition method. The experiment results show that the proposed statistical pattern-recognition method for driving styles based on kernel density estimation is more efficient and stable than the fuzzy logic-based method.