Search

Grid state estimation is essential for effective control and management of distribution systems. While weighted least squares has been the conventional method for state estimation, sparsity-aware methods have become popular due to their superior performance with limited data. Matrix completion and compressed sensing-based state estimation approaches exploit the underlying smoothness in the state variables. However, classic matrix completion methods do not take into account the temporal correlation of system states. Compressed sensing methods, on the other hand, require an appropriate choice of sparsifying basis that may not be easy to identify. This article proposes a block tensor completion based framework which uses an alternative approach to estimate voltage phasor, power injections, and branch currents. This approach utilizes the temporal correlation of the system states in a tensor trace-norm minimization formulation with power flow equations as constraints. Feature scaling is introduced in the problem formulation to benefit from the improved sensitivity of the tensor trace norm to the matrix columns in the scaled unfoldings of the tensor. Weighted tensor norm is utilized to exploit the structures of the different unfoldings of the state measurement tensor to improve the voltage estimation. The estimation accuracy is further improved by alternatively estimating the tensor columns and increasing the available data at each stage in the tensor completion process. The proposed methods are evaluated on the IEEE-33, 37 test systems, and a 100-node test system. The proposed methods are shown to provide significant performance gains relative to the classic matrix and tensor completion based approaches.

In this article, an automated heartbeat classification has been proposed to prevent the growing threats of cardiovascular diseases around the world. The MIT-BIH arrhythmia database has been used for the training and testing of the proposed approach. The database contains 90 095 normal (N), 2781 supraventricular (SVEB), 7008 ventricular (VEB), 802 fusions (F), and 15 unclassified (Q) beats each of 30 min duration. Total 61 features have been extracted using the time-series feature extraction library (TSFEL). Feature selection or reduction methods applied are feature scaling, removal of highly correlated and low variance features, and random forest (RF) recursive feature elimination. The methodological novelty of this study is mainly the incorporation of TSFEL during feature extractions, synthetic minority oversampling technique to create a balanced dataset, and an ensemble of RF and support vector machine (SVM) using weighted majority algorithm for heartbeat classification to improve the results. Grid search has been performed to optimize the hyperparameters of RF and SVM classifiers. A final evaluation has been carried out considering a “subject-specific” scheme. The sensitivity that our approach has achieved for the arrhythmic heartbeat classes is as follows: N: 99.50%, SVEB (S): 74.20%, VEB (V): 94.22%, F: 73.21%, and Q: 0%. The corresponding positive predictive values are N: 98.67%, SVEB (S): 90.09%, VEB (V): 95.95%, F: 88.35%, and Q: 0%. In comparison with machine learning and deep learning based state-of-the-art approaches, significant improvement in efficiency has been found. Impact Statement— Automated classifications of heartbeats from ECG signals are very much required to speed up the diagnosis of cardiovascular diseases (CVDs). Our study demonstrates 98.21% accuracy compared to existing state-of-the-art approaches. This study impacts the intensive health care system by offering an easy, quick, and cost-effective diagnosis of CVDs. Besides, remote patients can also receive the treatments timely. Early intervention and prevention of CVDs can prevent almost 30% of the total death every year. The proposed approach has a societal and economic impact too, as the health, welfare, and productivity of the citizens as well as of the country will be improved significantly. Approximately 2% of gross domestic product (GDP) will be improved across low- and middle-income countries as the proposed approaches will reduce the direct (treatment cost) and indirect costs (loss in productivities at workplaces) associated with CVDs.

Image denoising is a classical topic yet still a challenging problem, especially for reducing noise from the texture information. Feature scaling (e.g., downscale and upscale) is a widely practice in image denoising to enlarge receptive field size and save resources. However, such a common operation would lose some visual informative details. To address those problems, we propose fast and accurate image denoising via attention guided scaling (AGS). We find that the main informative feature channel and visual primitives during the scaling should keep similar. We then propose to extract the global channel-wise attention to maintain main channel information. Moreover, we propose to collect global descriptors by considering the entire spatial feature. And we then distribute the global descriptors to local positions of the scaled feature, based on their specific needs. We further introduce AGS for adversarial training, resulting in a more powerful discriminator. Extensive experiments show the effectiveness of our proposed method, where we clearly surpass all the state-of-the-art methods on most popular synthetic and real-world denoising benchmarks quantitatively and visually. We further show that our network contributes to other high-level vision applications and improves their performances significantly.

Multi-label classification aims to associate multiple labels to a given data/object instance to better describe them. Multi-label data sets are common in a lot of emerging application areas like: Text/Multimedia classification, Bio-Informatics, Medical image annotations and Computer Vision to name a few. There is a growing interest in efficient and accurate multi-label classification. There are two major approaches to perform multi-label classification (i) problem transformation methods and (ii) algorithm adaptation methods. In algorithm adaptation, the traditional classification algorithms are modified to handle multi-label data sets. One classification algorithm which is often modified to do multi-label classification is k- nearest neighbor (kNN). k-nearest neighbor is popular due to its simplicity, easy to implement and seamlessly adaptability. Despite its merits it has several drawbacks like: sensitivity of noisy data, missing values and outliers; feature scaling and often becoming inaccurate for large overlapping solution space. In this paper, a modification to kNN method is suggested for multi-label classification with three improvement strategies (i) selection of local example w.r.t. unknown example - the motivation for this comes from the fact that local and relevant space is vital for the improvement in multi-label classification; (ii) Splitting the input space into multiple sub-spaces for optimal label estimation - the motivation is to estimate label accurately in the presence of noisy labels; And (iii) selection of labels using Mean Average Precision (MAP) estimates - here our motivation is to utilize the training data effectively to maximize the hidden distribution and optimal parameters for the method. The proposed method is implemented and compared with state-of-the-art approaches based on kNN or similar approaches that effectively select and optimize relevant spaces for multi-label classification. Evaluation based on multiple metrics like Hamming loss, Precision/Recall and F-measure are used for evaluation. The suggested approach performed much better than the state-of-the-art on the datasets with strong label cardinalities.

Accurate load forecasting can create both economic and reliability benefits for power system operators. However, the cyberattack on load forecasting may mislead operators to make unsuitable operational decisions for the electricity delivery. To effectively and accurately detect these cyberattacks, this paper develops a machine learning-based anomaly detection (MLAD) methodology. First, load forecasts provided by neural networks are used to reconstruct the benchmark and scaling data by using the ${k}$ -means clustering. Second, the cyberattack template is estimated by the naive Bayes classification based on the cumulative distribution function and statistical features of the scaling data. Finally, the dynamic programming is utilized to calculate both the occurrence and parameter of one cyberattack on load forecasting data. A widely used symbolic aggregation approximation method is compared with the developed MLAD method. Numerical simulations on the publicly load data show that the MLAD method can effectively detect cyberattacks for load forecasting data with relatively high accuracy. Also, the robustness of MLAD is verified by thousands of attack scenarios based on Monte Carlo simulation.

In this paper, a received signal strength indicator (RSSI) based indoor localization system is implemented employing Wi-Fi infrastructure. In the light of the feature-scaling-based $k$ -nearest neighbor (FS- $k$ NN) algorithm, a new continuous-feature-scaling model is proposed, which uses continuous weights instead of the discrete weights used in the FS- $k$ NN, and needs not divide the entire RSSI space into the intervals. This gridless scheme avoids the difficulty of the weight selection at the common boundary of the adjacent intervals that could meet in the grid-based method of FS- $k$ NN, which needs to divide the RSSI space into the intervals, ahead. An outlier deleting procedure is further used to improve the accuracy of the localization system. Experimental results indicate that the proposed method can be with a small localization error and is superior to some previous methods. One of the experiments achieves 1.34 m of the indoor localization error in a 12 m $\times 8\,\,{\mathrm{m}}$ area. The other is with 1.72 m of the indoor localization error in a 30 m $\times 25\,\,{\mathrm{m}}$ area. The proposed method performances best among the counterparts in the experiments.

Scaling analysis of several thousand Nb 3 Sn critical-current (I c ) measurements is used to derive the extrapolative scaling expression (ESE), a fitting equation that can quickly and accurately extrapolate (or interpolate) limited datasets to obtain full three-dimensional dependences of I c on magnetic field (B), temperature (T), and mechanical strain (e). Unlike nonextrapolative fitting equations, the ESE relation is determined completely by fundamental raw scaling data from master pinning-force curves. The results show that extrapolation errors with ESE approach typical I c measurement errors. The scaling expression is simple and robust, providing straightforward extrapolation capability for conductor characterization and magnet design. The ESE relation offers the prospect for extrapolations in several new areas, including: a reduction in the measurement space required for full I c (B, T, e) characterization to about one fifth the size; ability to combine data from separate temperature and strain apparatuses; extrapolation of transport I c data from above 4 K to lower temperatures, where heating effects and instabilities are problematic for transport measurements; and extrapolation of full I c (B, T, e) datasets from as little as a single I c (B) curve when several core parameters have been measured in similar conductors (particularly applicable to qualifying production wires). Accuracies are evaluated for concatenations of these different extrapolation capabilities. Examples are given for practical Nb 3 Sn conductors, including those for high luminosity-LHC magnets, ITER, and cryocooled NMR magnets.

With the increasing popularity of WLAN infrastructure, WiFi fingerprint-based indoor positioning systems have received considerable attention recently. Much existing work in this aspect adopts classification techniques that match a vector of radio signal strengths (RSSs) reported by a mobile station (MS) to pretrained reference fingerprints sampled from different access points (APs) at different reference points (RPs) with known positions. However, in the calculation of signal distances between different RSS vectors, existing techniques fail to consider the fact that equal RSS differences at different RSS levels may not mean equal differences in geometrical distances in complex indoor environment. To address this issue, in this paper, we propose a feature-scaling-based $k$ -nearest neighbor (FS- $k$ NN) algorithm for achieving improved localization accuracy. In FS- $k$ NN, we build a novel RSS-level-based FS model, which introduces RSS-level-based scaling weights in the computation of effective signal distances between signal vector reported by a MS and reference fingerprints in a radio map. Experimental results show that FS- $k$ NN can achieve an average location error as low as 1.70 m, which is superior to existing work.

Assessment of mental workload using physiological measures, especially electroencephalography (EEG) signals, is an active area. Recently, a number of wireless acquisition systems to measure EEG and other physiological signals have become available. Few studies have applied such wireless systems to assess cognitive workload and evaluate their performance. This paper presents an initial step to explore the feasibility of a popular wireless system (Emotiv EPOC headset) to assess memory workload levels in a well-known $n$ -back task. We developed a signal processing and classification framework, which integrated an automatic artifact removal algorithm, a broad spectrum of feature extraction techniques, a personalized feature scaling method, an information-theory-based feature selection approach, and a proximal-support-vector-machine-based classification model. The experimental results show that the wirelessly collected EEG signals can be used to classify different memory workload levels for nine participants. The classification accuracies between the lowest workload level (0-back) and active workload levels (1-, 2-, 3-back) were close to 100%. The best classification accuracy for 1- versus 2-back was 80%, and 1- versus 3-back was 84%. This study indicates that the wireless acquisition system and the advanced data analytics and pattern recognition techniques are promising to achieve real-time monitoring and identification of mental workload levels for humans engaged in a wide variety of cognitive activities in the modern society.

Feature selection is often necessary when implementing classifiers in practice. Most approaches to feature selection are motivated by the curse of dimensionality, but few seek to mitigate the overall computational cost of feature extraction. In this work, we propose a model-based approach for addressing both objectives. The model is based around a sparse kernel machine with feature scaling parameters controlled by a beta-Bernoulli prior. The hyperparameters are controlled by each feature’s computational cost. Experiments were carried out using publicly-available data sets, and the proposed Cost-Constrained Feature Optimization (CCFO) was compared to related methods in terms of accuracy and computational reduction.

Dental restoration begins with removing carries and affected tissues with air-turbine rotary cutting handpieces, and later restoring the lost tissues with appropriate restorative materials to retain the functionality. Most restoration materials eventually fail as they age and need to be replaced. One of the difficulties in replacing failing restorations is discerning the boundary of restorative materials, which causes inadvertent removal of healthy tooth layers. Developing an objective and sensor-based method is a promising approach to monitor dental restorative operations and to prevent excessive tooth losses. This paper has analyzed cutting sounds of an air-turbine handpiece to discriminate between tooth layers and two commonly used restorative materials, amalgam and composite. Support vector machines were employed for classification, and the averaged short-time Fourier transform coefficients were selected as the features. The classifier performance was evaluated from different aspects such as the number of features, feature scaling methods, classification schemes, and utilized kernels. The total classification accuracies were 89% and 92% for cases included composite and amalgam materials, respectively. The obtained results indicated the feasibility and effectiveness of the proposed method.

This paper proposes an unsupervised discrimination analysis for feature selection based on a property of the Fourier transform of the probability density distribution. Each feature is evaluated on the basis of a simple observation motivated by the concept of optical diffraction, which is invariant under feature scaling. The time complexity is O(mn), where m is number of features and n is number of instances when being applied directly to the given data. This approach is also extended to deal with data orientation, which is the direction of data alignment. Therefore, the discrimination score of any transformed space can be used for evaluating the original features. The experimental results on several real-world datasets demonstrate the effectiveness of the proposed method.

We consider projected equations for approximate solution of high-dimensional fixed point problems within low-dimensional subspaces. We introduce an analytical framework based on an equivalence with variational inequalities, and algorithms that may be implemented with low-dimensional simulation. These algorithms originated in approximate dynamic programming (DP), where they are collectively known as temporal difference (TD) methods. Even when specialized to DP, our methods include extensions/new versions of TD methods, which offer special implementation advantages and reduced overhead over the standard LSTD and LSPE methods, and can deal with near singularity in the associated matrix inversion. We develop deterministic iterative methods and their simulation-based versions, and we discuss a sharp qualitative distinction between them: the performance of the former is greatly affected by direction and feature scaling, yet the latter have the same asymptotic convergence rate regardless of scaling, because of their common simulation-induced performance bottleneck.

When applying gradient descent learning methods to hierarchical fuzzy systems, there is great difficulty in handling the intermediate variables introduced by the hierarchical structures, as the intermediate variables may go outside their definition domain that makes gradient descent learning invalid. To overcome this difficulty, this paper proposes a learning scheme that integrates a normalization process for intermediate variables into gradient descent learning. This ensures that gradient descent methods are applicable to, and correctly used for, learning general hierarchical fuzzy systems. Benchmark datasets are used to demonstrate the validity and advantages of the proposed learning scheme over other existing methods in terms of better accuracy, better transparency, and fewer fuzzy rules and parameters.