Statistical Pattern Recognition Approach Research Articles

Model-informed deep learning strategy with vision measurement for damage identification of truss structures

Structural damage identification approaches can be divided into two categories, i.e. data-driven approaches via statistical pattern recognition and model-based approaches via finite element (FE) model updating. These two approaches have their own merits, and their main shortcomings can be remedied by each other’s merits. Therefore, this study proposed a deep learning-based damage identification strategy involving both data-driven and model-based approaches, termed as model-informed deep learning (MIDL)-based strategy. This strategy first proposes a vision-based displacement estimation approach to extract structural displacement responses from video data. This approach reduces the displacement drift induced by conventional optical flow approaches and improves the tracking accuracy of feature points. Then, a calibrated FE model is built to construct data sets with different damage levels via FE model updating and time-history analysis. Following this, a one-dimensional convolutional neural network (1D CNN) is established to detect and localize structural damage by using direct displacement responses. Finally, FE model updating is performed again to quantify structural damage level with constrained targets. A truss structure is further used to evaluate the accuracy of the proposed strategy experimentally. Results illustrate that the proposed MIDL strategy, which uses time series to localize the structural damage, achieves a global location accuracy of 86.09% and avoids the feature extraction process. Meanwhile, with the known damage location, the efficiency and accuracy of structural damage quantification via rerunning model updating can also be significantly improved. In addition, the displacements estimated by the proposed approach have a good match with ground truth values with error standard deviations of less than 0.3 mm.

An Empirical Study on IPO Model Construction of Undergraduate Education Quality Evaluation in China from the Statistical Pattern Recognition Approach In NLP

Based on the analysis of 1,497 samples from the survey of national undergraduate educational administrators, the IPO model of undergraduate education quality evaluation from the perspective of managers can be effectively verified. The quality of higher education needs the accountability of higher education and evaluation of student learning outcomes. The empirical studies show the effects on the various dimensions of quality provisions which were not the same. The main findings are that the input of undergraduate education and teaching can not only directly and positively predict the output of undergraduate education and teaching, but also can positively predict the output of undergraduate education and teaching through the partial mediating effect of the process of undergraduate education and teaching. The research suggests that under the given material conditions, it is essential to enhance the “soft input” of undergraduate education and teaching, to strengthen the process construction of undergraduate education and teaching, to enhance the process quality control in the implementation of humanistic care, to pay attention to the development effect of teachers and students and the development effect of school-running characteristics, and to promote the connotative development of colleges and universities. The study provides a practical framework, model, and guidelines that can be used for undergraduate education institutions to evaluate and enhance the performance to effectively work in society. The quality evaluation of undergraduate education needs to focus from the quality of students' learning outcomes to the comprehensive consideration of “input-process-outcome”.

Diagnostic of fuel cell air supply subsystems based on pressure signal records and statistical pattern recognition approach

A data-driven and application-oriented diagnosis tool is developed for Fuel Cell (FC) air supply subsystems. A bench emulating a FC air line is built to study normal and abnormal operations (clogged inlet, air leakage, error in compressor speed control) and data are collected using the air pressure transducer, which is usually implemented in FC generators. A pattern recognition approach is then applied to statistical features extracted from the pressure signal. The performance of the diagnosis strategy is evaluated from confusion matrices, associated to graphs and performance indicators. Two examples of compressors, air subsystem managements, and data records are considered to examine the method portability. Best classification rates (>95%) are obtained on test profiles, when the pressure regulation is disabled; fault stamps can thus be found in the pressure signal morphology. Regarding the frequency of data logging, both 1 kHz and 100 Hz values are found effective for fault isolations.

Fusion of smartphone sensor data for classification of daily user activities

New mobile applications need to estimate user activities by using sensor data provided by smart wearable devices and deliver context-aware solutions to users living in smart environments. We propose a novel hybrid data fusion method to estimate three types of daily user activities (being in a meeting, walking, and driving with a motorized vehicle) using the accelerometer and gyroscope data acquired from a smart watch using a mobile phone. The approach is based on the matrix time series method for feature fusion, and the modified Better-than-the-Best Fusion (BB-Fus) method with a stochastic gradient descent algorithm for construction of optimal decision trees for classification. For the estimation of user activities, we adopted a statistical pattern recognition approach and used the k-Nearest Neighbor (kNN) and Support Vector Machine (SVM) classifiers. We acquired and used our own dataset of 354 min of data from 20 subjects for this study. We report a classification performance of 98.32 % for SVM and 97.42 % for kNN.

Age Prediction using Image Dataset using Machine Learning

Gender is a central feature of our personality still. In our social life it is also an significant element. Artificial intelligence age predictions can be used in many fields, such as smart human-machine interface growth , health, cosmetics, electronic commerce etc. The prediction of people's sex and age from their facial images is an ongoing and active problem of research. The researchers suggested a number of methods to resolve this problem, but the criteria and actual performance are still inadequate. A statistical pattern recognition approach for solving this problem is proposed in this project.Convolutionary Neural Network (ConvNet / CNN), a Deep Learning algorithm, is used as an extractor of features in the proposed solution. CNN takes input images and assigns value to different aspects / objects (learnable weights and biases) of the image and can differentiate between them. ConvNet requires much less preprocessing than other classification algorithms. While the filters are hand-made in primitive methods, ConvNets can learn these filters / features with adequate training.In this research, face images of individuals have been trained with convolutionary neural networks, and age and sex with a high rate of success have been predicted. More than 20,000 images are containing age, gender and ethnicity annotations. The images cover a wide range of poses, facial expression, lighting, occlusion, and resolution.

Driving-Style-Oriented Adaptive Equivalent Consumption Minimization Strategies for HEVs

The performance of energy management systems in hybrid electric vehicles (HEVs) is highly related to drivers’ driving style. This paper proposes a driving-style-oriented adaptive equivalent consumption minimization strategy (AECMS-style) in order to improve fuel economy for HEVs. For this purpose, first, a statistical pattern recognition approach is proposed to classify drivers into six groups from moderate to aggressive using kernel density estimation and entropy theory. Then, the effects of driving style on energy management strategies are discussed by analyzing the performance of the equivalent consumption minimization strategy (ECMS). Based on the comprehensive analysis, we design a new optimal equivalent factor adjustment rule for the AECMS-style and also redesign the braking strategy of motors at driving charging mode for different driving styles. Finally, five drivers with typical driving styles participate in experiments to show the effectiveness of our proposed method. Experimental results demonstrate that the AECMS-style can improve the fuel economy and charging sustainability of HEVs, compared with ECMS.

Statistical pattern recognition approach for long-time monitoring of the G.Meazza stadium by means of AR models and PCA

In recent years, the interest for the automatic evaluation of the state of civil structures is increased. The development of Structural Health Monitoring is allowed by the low costs of the hardware and the improving of the computational capacity of computers that can analyze considerable amount of data in short time. A Structural Health Monitoring (SHM) system should continuously monitor structures, extracting and processing relevant information, to efficiently allocate the resources for maintenance and ensure the security of the structure. Considering the latest developments in this field, great attention has been paid to data-based approaches, especially to autoregressive models; these econometric models, born in the field of finance, are usually used to analyze the vibration time series provided by the sensors applied on the monitored structures. Indexes based on these autoregressive models can be used as features by which the structural integrity can be assessed. This work proposes the application of a multivariable analysis, Principal Component Analysis (PCA), to the set of the autoregressive model parameters estimated on the vibration responses of a real structure under operational conditions. This approach reduces a complex set of data to a lower dimension, by representing the behavior of the structure through the few variables. This work uses the principal components of the autoregressive model parameters as indicators that can effectively describe different operational levels and some important environmental effects. The strategy is applied for the first time on the data collected by the long-time monitoring system installed on the stands of the G. Meazza stadium in Milan. The results will show that this procedure is effective in representing the status of the structure and can be used in a structural health monitoring prospective.

Performance Analysis of Statistical Pattern Recognition Methods in KEEL

This article represents the focused performance analysis of the statistical Pattern Recognition methods. There is a variety of emerging pattern recognition problems for different data representations like multimedia, spatial, temporal and textual data. Statistical Pattern Recognition approaches are extensively applied for pattern recognition and classification purposes. There are various statistical pattern-recognition methods proposed in the contemporary literature. For the selection of more appropriate statistical pattern recognition, method demands comprehensive performance evaluation of contemporary methods. However, this empirical study is focused on the performance evaluation of statistical methods for pattern discovery in terms of accuracy and error rate. The studied methods include Naïve-Bayes (NB-C), Linear Discriminant Analysis (LDA-C), Kernel Classifier (Kernel-C), Least Mean Square Linear Classifier (LinearLMS-C), Least Mean Square Quadratic classifier(PolQuadraticLMS-C), Multinomial logistic regression model with a ridge estimator(Logistic-C) and Particle Swarm Optimization - Linear Discriminant Analysis (PSOLDA-C). The implementation of mentioned methods in KEEL, a data mining tool, is applied on public datasets for the comparative analysis. Experimental results reveal that the performance of PSOLDA-C is promising than other methods in terms of accuracy.

Data-based structural health monitoring using small training data sets

One of the most efficient ways to solve the damage detection problem using the statistical pattern recognition approach is that of exploiting the methods of outlier analysis. Cast within the pattern recognition framework, damage detection assesses whether the patterns of the damage-sensitive features extracted from the response of the system under unknown conditions depart from those drawn by the features extracted from the response of the system in a healthy state. The metric dominantly used to measure the testing feature's departure from the trained model is the Mahalanobis Squared Distance (MSD). Evaluation of the MSD requires the use of the inverse of the training population's covariance matrix. It is known that when the feature dimensions are comparable with the number of observations, the covariance matrix is ill-conditioned and numerically problematic to invert. When the number of observations is smaller than the feature dimensions, the covariance matrix is not even invertible. In this paper, four alternatives to the canonical damage detection procedure are investigated to address this issue: data compression through discrete cosine transform, use of pseudo-inverse of the covariance matrix, use of shrinkage estimate of the covariance matrix, and a combination of the three aforementioned techniques. The performance of the four methods is first studied on simulated data and then compared using the experimental data recorded on a four-story steel frame excited at the base by means of a shaking table available at the Carleton Laboratory at Columbia University. Copyright © 2015 John Wiley & Sons, Ltd.

Combining Multiple Dynamic Models and Deep Learning Architectures for Tracking the Left Ventricle Endocardium in Ultrasound Data

We present a new statistical pattern recognition approach for the problem of left ventricle endocardium tracking in ultrasound data. The problem is formulated as a sequential importance resampling algorithm such that the expected segmentation of the current time step is estimated based on the appearance, shape, and motion models that take into account all previous and current images and previous segmentation contours produced by the method. The new appearance and shape models decouple the affine and nonrigid segmentations of the left ventricle to reduce the running time complexity. The proposed motion model combines the systole and diastole motion patterns and an observation distribution built by a deep neural network. The functionality of our approach is evaluated using a dataset of diseased cases containing 16 sequences and another dataset of normal cases comprised of four sequences, where both sets present long axis views of the left ventricle. Using a training set comprised of diseased and healthy cases, we show that our approach produces more accurate results than current state-of-the-art endocardium tracking methods in two test sequences from healthy subjects. Using three test sequences containing different types of cardiopathies, we show that our method correlates well with interuser statistics produced by four cardiologists.

Salient and non-salient fiducial detection using a probabilistic graphical model

Deformable shape detection is an important problem in computer vision and pattern recognition. However, standard detectors are typically limited to locating only a few salient landmarks such as landmarks near edges or areas of high contrast, often conveying insufficient shape information. This paper presents a novel statistical pattern recognition approach to locate a dense set of salient and non-salient landmarks in images of a deformable object. We explore the fact that several object classes exhibit a homogeneous structure such that each landmark position provides some information about the position of the other landmarks. In our model, the relationship between all pairs of landmarks is naturally encoded as a probabilistic graph. Dense landmark detections are then obtained with a new sampling algorithm that, given a set of candidate detections, selects the most likely positions as to maximize the probability of the graph. Our experimental results demonstrate accurate, dense landmark detections within and across different databases.

A statistical pattern recognition approach for the classification of cooking stages. The boiling water case

Although pattern recognition technique has been widely used in many fields, it seems that very few studies have applied this technique to cooking processes. In this preliminary research, a new methodology has been developed and tested on a simple case of water boiling. Besides defining and analysing the efficacy and the performance of a statistical pattern recognition approach when applied to different signals (sound and vibration), an optimisation module has been proposed to boost the classification rates by adding syntactical analysis that enables the inertia of the process to be considered. In the specific case of boiling water, almost 100% successful recognition has been reached. These results prove the validity of this methodology, opening up new research lines for new scenarios, such as different cooking processes, acoustically polluted environments, and sensor optimisation.

PERANCANGAN SISTEM PAKAR UNTUK IDENTIFIKASI DAN REKOMENDASI PERBAIKAN KERUSAKAN PADA MOTOR AC

Normal 0 false false false MicrosoftInternetExplorer4 Abstract : The design of expert system for identifying and recommending fault recovery of AC motor. The existing vibration measurement instruments do not have capability to identify AC motor fault and give the information about the cause of it and reparation recommendation. Therefore, in this research, it will design the AC motor fault identification system through measurement data. The causes of motor damage that can be identified are short lifetime bearing, over vibration, and bearing overheating. The identification process uses the statistical pattern recognition approach. From this approach, it’s obtained reliable interval 0.0247 to 0.1486 V for short lifetime bearing, 0.5022 to 0.705 V for over vibration, and 0.7183 to 0.9812 V for bearing overheating. From this test, it’s obtained the reliable level of this system approximately 80%. /* Style Definitions */ table.MsoNormalTable {mso-style-name:Table Normal; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-parent:; mso-padding-alt:0cm 5.4pt 0cm 5.4pt; mso-para-margin:0cm; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:10.0pt; font-family:Times New Roman; mso-ansi-language:#0400; mso-fareast-language:#0400; mso-bidi-language:#0400;}

Automatic Speech Recognition Based on Non-Uniform Error Criteria

The Bayes decision theory is the foundation of the classical statistical pattern recognition approach, with the expected error as the performance objective. For most pattern recognition problems, the “error” is conventionally assumed to be binary, i.e., 0 or 1, equivalent to error counting, independent of the specifics of the error made by the system. The term “error rate” is thus long considered the prevalent system performance measure. This performance measure, nonetheless, may not be satisfactory in many practical applications. In automatic speech recognition, for example, it is well known that some errors are more detrimental (e.g., more likely to lead to misunderstanding of the spoken sentence) than others. In this paper, we propose an extended framework for the speech recognition problem with non-uniform classification/recognition error cost which can be controlled by the system designer. In particular, we address the issue of system model optimization when the cost of a recognition error is class dependent. We formulate the problem in the framework of the minimum classification error (MCE) method, after appropriate generalization to integrate the class-dependent error cost into one consistent objective function for optimization. We present a variety of training scenarios for automatic speech recognition under this extended framework. Experimental results for continuous speech recognition are provided to demonstrate the effectiveness of the new approach.

A modified statistical pattern recognition approach to measuring the crosslinguistic similarity of Mandarin and English vowels

This study describes a statistical approach to measuring crosslinguistic vowel similarity and assesses its efficacy in predicting L2 learner behavior. In the first experiment, using linear discriminant analysis, relevant acoustic variables from vowel productions of L1 Mandarin and L1 English speakers were used to train a statistical pattern recognition model that simultaneously comprised both Mandarin and English vowel categories. The resulting model was then used to determine what categories novel Mandarin and English vowel productions most resembled. The extent to which novel cases were classified as members of a competing language category provided a means for assessing the crosslinguistic similarity of Mandarin and English vowels. In a second experiment, L2 English learners imitated English vowels produced by a native speaker of English. The statistically defined similarity between Mandarin and English vowels quite accurately predicted L2 learner behavior; the English vowel elicitation stimuli deemed most similar to Mandarin vowels were more likely to elicit L2 productions that were recognized as a Mandarin category; English stimuli that were less similar to Mandarin vowels were more likely to elicit L2 productions that were recognized as new or emerging categories.

Sound Categories Are Represented as Distributed Patterns in the Human Auditory Cortex

The ability to recognize sounds allows humans and animals to efficiently detect behaviorally relevant events, even in the absence of visual information. Sound recognition in the human brain has been assumed to proceed through several functionally specialized areas, culminating in cortical modules where category-specific processing is carried out. In the present high-resolution fMRI experiment, we challenged this model by using well-controlled natural auditory stimuli and by employing an advanced analysis strategy based on an iterative machine-learning algorithm that allows modeling of spatially distributed, as well as localized, response patterns. Sounds of cats, female singers, acoustic guitars, and tones were controlled for their time-varying spectral characteristics and presented to subjects at three different pitch levels. Sound category information--not detectable with conventional contrast-based methods analysis--could be detected with multivoxel pattern analyses and attributed to spatially distributed areas over the supratemporal cortices. A more localized pattern was observed for processing of pitch laterally to primary auditory areas. Our findings indicate that distributed neuronal populations within the human auditory cortices, including areas conventionally associated with lower-level auditory processing, entail categorical representations of sounds beyond their physical properties.

A statistical pattern recognition approach for determining cellular viability and lineage phenotype in cultured cells and murine bone marrow

Cellular binding of annexin V and membrane permeability to 7-aminoactinomycin D (7AAD) are important tools for studying apoptosis and cell death by flow cytometry. Combining viability markers with cell surface marker expression is routinely used to study various cell lineages. Current classification methods using strict thresholds, or "gates," on the fluorescent intensity of these markers are subjective in nature and may not fully describe the phenotypes of interest. We have developed objective criteria for phenotypic boundary recognition through the application of statistical pattern recognition. This task was achieved using artificial neural networks (ANNs) that were trained to recognize subsets of cells with known phenotypes, and then used to determine decision boundaries based on statistical measures of similarity. This approach was then used to test the hypothesis that erythropoietin (EPO) inhibits apoptosis and cell death in erythroid precursor cells in murine bone marrow. Our method was developed for classification of viability using an in vitro cell system and then applied to an ex vivo analysis of murine late-stage erythroid progenitors. To induce apoptosis and cell death in vitro, an EPO-dependent human leukemic cell line, UT-7(EPO) cells were incubated without recombinant human erythropoietin (rhEPO) for 72 h. Five different ANNs were trained to recognize live, apoptotic, and dead cells using a "known" subset of the data for training, and a K-fold cross validation procedure for error estimation. The ANNs developed with the in vitro system were then applied to classify cells from an ex vivo study of rhEPO treated mice. Tg197 (human tumor necrosis-alpha transgenic mice, a model of anemia of chronic disease) received a single s.c. dose of 10,000 U/kg rhEPO and femoral bone marrow was collected 1, 2, 4, and 8 days after dosing. Femoral bone marrow cells were stained with TER-119 PE, CD71 APC enable identification of erythroid precursors, and annexin V FITC and 7AAD to identify the apoptotic and dead cells. During classification forward and side angle light scatter were also input to all pattern recognition systems. Similar decision boundaries between live, apoptotic, and dead cells were consistently identified by the neural networks. The best performing network was a radial basis function multi-perceptron that produced an estimated average error rate of 4.5% +/- 0.9%. Using these boundaries, the following results were reached: depriving UT-7(EPO) cells of rhEPO induced apoptosis and cell death while the addition of rhEPO rescued the cells in a dose-dependent manner. In vivo, treatment with rhEPO resulted in an increase of live erythroid cells in the bone marrow to 119.8% +/- 9.8% of control at the 8 day time point. However, a statistically significant transient increase in TER-119(+) CD71(+) 7AAD(+) dead erythroid precursors was observed at the 1 and 2 day time points with a corresponding decrease in TER-119(+) CD71(+) 7AAD(-) Annexin V(-) live erythroid precursors, and no change in the number of TER-119(+) CD71(+) annexin V(+) 7AAD(-) apoptotic erythroid precursors in the bone marrow. A statistical pattern recognition approach to viability classification provides an objective rationale for setting decision boundaries between "positive" and "negative" intensity measures in cytometric data. Using this approach we have confirmed that rhEPO inhibits apoptosis and cell death in an EPO dependent cell line in vitro, but failed to do so in vivo, suggesting EPO may not act as a simple antiapoptotic agent in the bone marrow. Rather, homeostatic mechanisms may regulate the pharmacodynamic response to rhEPO.

Sensor network paradigms for structural health monitoring

Structural health monitoring (SHM) is the process of detecting damage in structures. The goal of SHM is to improve the safety and reliability of aerospace, civil and mechanical infrastructure by detecting damage before it reaches a critical state. A specific topic that has not been extensively addressed in the SHM literature is the development of rigorous approaches to designing the SHM sensing system that is used to address the data acquisition portion of the problem. To date, almost all such system designs are done somewhat in an ad hoc manner where the engineer picks a sensing system that is readily available and that they are familiar with, and then attempts to demonstrate that a specific type of damage can be detected with that system. In many cases this approach has been shown to be ineffective and as a result researchers have begun to develop sensor networks specially suited for SHM. Based on this research, several sensor network paradigms for SHM have emerged, and this paper is intended to provide an overview of these paradigms. This paper will first provide a brief summary of the statistical pattern recognition approach to SHM problem. The data acquisition portion of the paradigm is then addressed in detail where the various parameters of the system that must be considered in its design and subsequent field deployment are summarized. Published in 2005 by John Wiley & Sons, Ltd.

Identifying estrogen receptor target genes using integrated computational genomics and chromatin immunoprecipitation microarray

The estrogen receptor alpha (ERalpha) regulates gene expression by either direct binding to estrogen response elements or indirect tethering to other transcription factors on promoter targets. To identify these promoter sequences, we conducted a genome-wide screening with a novel microarray technique called ChIP-on-chip. A set of 70 candidate ERalpha loci were identified and the corresponding promoter sequences were analyzed by statistical pattern recognition and comparative genomics approaches. We found mouse counterparts for 63 of these loci and classified 42 (67%) as direct ERalpha targets using classification and regression tree (CART) statistical model, which involves position weight matrix and human-mouse sequence similarity scores as model parameters. The remaining genes were considered to be indirect targets. To validate this computational prediction, we conducted an additional ChIP-on-chip assay that identified acetylated chromatin components in active ERalpha promoters. Of the 27 loci upregulated in an ERalpha-positive breast cancer cell line, 20 having mouse counterparts were correctly predicted by CART. This integrated approach, therefore, sets a paradigm in which the iterative process of model refinement and experimental verification will continue until an accurate prediction of promoter target sequences is derived.

The Use of Vibration Data for Damage Detection in Bridges: A Comparison of System Identification and Pattern Recognition Approaches

This paper briefly outlines the rationale for structural health monitoring as an integral component of bridge management systems. Two different approaches, system identification and statistical pattern recognition, are summarised and applied in turn to vibration data collected from three scale modelreinforced concrete bridges. The results show that the system identification paradigm can successfully locate and quantify the damage to the decks when they are loaded to incipient collapse, especially when experience is used to determine the parameters to use in the finite element updating procedure. However, the study also demonstrated that this approach requires a large amount of high quality data, requirements that cannot always be met readily in the field. In contrast, although the statistical pattern recognition approach was not able to quantify or locate the damage, it was able to clearly indicate that damage had occurred from relatively few measurements. A comparison of the strengths and weaknesses of the two approaches suggests that they should be used in a complementary manner. The statistical pattern recognition approach can be employed as a simple, cost efficient way to indicate that damage has occurred. It can then trigger a more detailed investigation using system identification.