Bearings are critical components in induction motors and brushless direct current motors. Bearing failure is the most common mode these By implementing health monitoring fault diagnosis of bearings, unscheduled maintenance economic losses caused by bearing failures can be avoided. This paper introduces trace ratio linear discriminant analysis (TR-LDA) to deal with high-dimensional non-Gaussian data for dimension reduction classification. Motor single-point faults generalized-roughness used...

In this paper, a new bearing anomaly detection and fault prognosis method is proposed. The detects anomalies then predicts its remaining useful life (RUL). To achieve these two goals, an autoregressive model, which used to filter out fault-unrelated signals, derived according healthy vibrational signals. A health index developed indicate conditions. Anomalies of bearings are detected by choosing appropriate threshold with the aid Box-Cox transformation. nonlinear model built track bearings'...

In this paper, we present new models and algorithms for object-level video advertising. A framework that aims to embed content-relevant ads within a stream is investigated in context. First, comprehensive optimization model designed minimize intrusiveness viewers when are inserted video. For human clothing advertising, design deep convolutional neural network using face features recognize genders stream. Human parts alignment then implemented extract part used retrieval. Second, develop...

Recently, domain adaptation has received extensive attention for solving intelligent fault diagnosis problems. It aims to reduce the distribution discrepancy between source and target through learning domain-invariant features. However, most existing methods mainly focus on global overlook subdomain adaptation, which results in loss of fine-grained information discriminative To address this problem, article, a deep adversarial network is proposed. This aligns relevant distributions...

In this paper, the authors present a real-time learning control scheme for unknown nonlinear dynamical systems using recurrent neural networks (RNNs). Two RNNs, based on same network architecture, are utilized in system. One is used to approximate system, and other mimic desired system response output. The rule achieved by combining two RNNs form A generalized iterative algorithm developed train RNNs. derived means of two-dimensional (2-D) theory that different from conventional algorithms...

This paper presents a novel technique for electric load forecasting based on neural weather compensation. Our proposed method is nonlinear generalization of Box and Jenkins approach nonstationary time-series prediction. A compensation network implemented one-day ahead forecasting. can accurately predict the change actual consumption from previous day. The results, Hong Kong Island historical demand, indicate that this methodology capable providing more accurate forecast with 0.9% reduction in error.

A novel framework using a Bayesian approach for content-based phishing web page detection is presented. Our model takes into account textual and visual contents to measure the similarity between protected suspicious pages. text classifier, an image algorithm fusing results from classifiers are introduced. An outstanding feature of this paper exploration estimate matching threshold. This required in classifier determining class identifying whether or not. In naive Bayes rule used calculate...

The tree structure is one of the most powerful structures for data organization. An efficient learning framework transforming tree-structured into vectorial representations presented. First, in attempting to uncover global discriminative information child nodes hidden at same level all trees, a clustering technique can be adopted allocating children different clusters, which are used formulate components vector. Moreover, locality-sensitive reconstruction method introduced model process,...

While the majority of methods for community detection produce disjoint communities nodes, most real-world networks naturally involve overlapping communities. In this paper, a scalable method in large is proposed. The based on an extension notion normalized cut to cope with A spectral clustering algorithm formulated solve related minimization problem. When available, may take into account prior information about likelihood each node belong several This can either be extracted from available...

Convolutional Neural Networks (CNNs) have demonstrated promising effectiveness in vibration-based fault diagnosis. However, the faulty characteristics are usually distributed on different scales and contaminated by noises from various sources. Therefore, it is still a challenging task for traditional CNNs to efficiently extract multiscale features suppress unrelated vibrational signals. In this paper, novel diagnosis framework called residual attention CNN (MRA-CNN) proposed learn...

A neural-based crowd estimation system for surveillance in complex scenes at underground station platform is presented. Estimation carried out by extracting a set of significant features from sequences images. Those feature indexes are modeled neural network to estimate the density. The learning phase based on our proposed hybrid least-squares and global search algorithms which capable providing characteristic fast convergence speed. Promising experimental results obtained terms accuracy...

A radial-basis-function (RBF) neural-network-based fault detection system is developed for performing induction machine and analysis. Four feature vectors are extracted from power spectra of vibration signals. The features inputs an RBF-type neural network identification classification. optimal architecture the RBF determined automatically by our proposed cell-splitting grid algorithm. This facilitates conventional laborious trial-and-error procedure in establishing architecture. In this...

A novel feature selection method using the concept of mutual information (MI) is proposed in this paper. In all MI based methods, effective and efficient estimation high-dimensional crucial. paper, a pruned Parzen window estimator quadratic (QMI) are combined to address problem. The results show that approach can estimate an way. With contribution, developed identify salient features one by one. Also, appropriate subsets for classification be reliably estimated. methodology thoroughly tested...

A wavelet transform based method was developed for diagnosing machine faults operating at different rotating speeds. This paper shows that fault diagnosis can be effectively performed when an appropriate narrow-band filter is used to extract the required spectra components. wavelets-transform-based technique design specified narrow banks. enables effective diagnostic analysis in frequency domain. Gaussian-enveloped oscillation-type employed. By matching basis functions with associated faulty...

This paper presents a Mahalanobis distance (MD) based diagnostic approach that employs probabilistic to establish thresholds classify product as being healthy or unhealthy. A technique for detecting trends and biasness in system health is presented by constructing control chart the MD value. The performance parameters' residuals, which are differences between estimated values (from an empirical model) observed monitoring), used isolate parameters exhibit faults. To aid qualification of...

We assess whether reactive oxygen species production and resistance to oxidative stress might be causally involved in the exceptional longevity exhibited by ocean quahog Arctica islandica. tested this hypothesis comparing production, stress, antioxidant defenses, protein damage elimination processes long-lived A islandica with shorter-lived hard clam, Mercenaria mercenaria. compared baseline biochemical profiles, age-related changes, responses exposure stressor tert-butyl hydroperoxide...

In this paper, we propose an efficient strategy to enhance traffic capacity via the process of nodes and links increment. We show that by adding shortcut existing networks, packets are avoided flowing through hub nodes. investigate performances our proposed under shortest path routing local strategy. Our obtained results using strategy, can be effectively enhanced Under is only when more likely forwarded low-degree in their paths. Compared with other strategies, indicate most effective...

Visualizing similarity data of different objects by exhibiting more separate organizations with local and multimodal characteristics preserved is important in multivariate analysis. Laplacian Eigenmaps (LAE) Locally Linear Embedding (LLE) aim at preserving the embeddings all pairs close vicinity reduced output space, but they are unable to identify interclass neighbors. This paper considers semi-supervised manifold learning problems. We apply pairwise Cannot-Link Must-Link constraints...

A new graph based constrained semi-supervised learning (G-CSSL) framework is proposed. Pairwise constraints (PC) are used to specify the types (intra- or inter-class) of points with labels. Since number labeled data typically small in SSL setting, core idea this create and enrich PC sets using propagated soft labels from both unlabeled by special label propagation (SLP), hence obtaining more supervised information for delivering enhanced performance. We also propose a Two-stage Sparse...

Isomap is a well-known nonlinear dimensionality reduction (DR) method, aiming at preserving geodesic distances of all similarity pairs for delivering highly manifolds. efficient in visualizing synthetic data sets, but it usually delivers unsatisfactory results benchmark cases. This paper incorporates the pairwise constraints into and proposes marginal (M-Isomap) manifold learning. The Cannot-Link Must-Link are used to specify types neighborhoods. M-Isomap computes shortest path over...