Search results

The purpose of this paper is to derive crucial insights from multi-scale analysis to detect equity return co-movements between Korean and emerging Asian markets.

Wavelet correlation, wavelet coherence and wavelet clustering measures are used to uncover Korean equity market interactions which are hard to see using any other modern econometric method and which would otherwise had remained hidden.

The authors observed that Korean equity market is strongly integrated with Asian equity markets at lower frequency scales and has a relatively weak correlation at higher frequencies. Further this correlation eventually grows strong in the interim of crises period at lower frequency scales. The authors, however, do not found any significant deviation in dendrograms generated in data clustering process from wavelet scale 2 to 6 which are associated with four and 64 weeks period, respectively. Overall the findings are relevant and have strong policy and practical implications.

The unique contribution of this paper is that it introduces wavelet clustering analysis to produce a nested hierarchy of similar markets at each frequency level for the first time in finance literature

In clinical analysis, medical image segmentation is an important step to study the anatomical structure. This helps to diagnose and classify abnormality in the image. The wide variations in the image modality and limitations in the acquisition process of instruments make this segmentation challenging. This paper aims to propose a semi-automatic model to tackle these challenges and to segment medical images.

The authors propose Legendre polynomial-based active contour to segment region of interest (ROI) from the noisy, low-resolution and inhomogeneous medical images using the soft computing and multi-resolution framework. In the first phase, initial segmentation (i.e. prior clustering) is obtained from low-resolution medical images using fuzzy C-mean (FCM) clustering and noise is suppressed using wavelet energy-based multi-resolution approach. In the second phase, resultant segmentation is obtained using the Legendre polynomial-based level set approach.

The proposed model is tested on different medical images such as x-ray images for brain tumor identification, magnetic resonance imaging (MRI), spine images, blood cells and blood vessels. The rigorous analysis of the model is carried out by calculating the improvement against noise, required processing time and accuracy of the segmentation. The comparative analysis concludes that the proposed model withstands the noise and succeeds to segment any type of medical modality achieving an average accuracy of 99.57%.

The proposed design is an improvement to the Legendre level set (L2S) model. The integration of FCM and wavelet transform in L2S makes model insensitive to noise and intensity inhomogeneity and hence it succeeds to segment ROI from a wide variety of medical images even for the images where L2S failed to segment them.

This paper describes fabric inspection system aided by computer vision to detect and classify defects in circular knitted fabrics using different common texture-recognition methods, including co-occurrence matrices, the discrete Fourier transform, wavelets, Gabor, and clustering. The images of the fabrics were broadly classified into six classes: cracks, holes, vertical stripes, horizontal stripes, soil freckles, and defect-free. One hundred and twenty images (256 gray level and 100 dpi) containing 20 images of defect-free fabrics (rib 1x1) as well as 100 images corresponding to five different categories were used. In general, one-half of the images in each category were employed for training and the remaining images were used for testing.

The application of the clustering method applied in this work was found to be highly promising at identifying defects in knitted fabrics. With an overall success rate of 91.6%, the clustering method has a higher efficiency value than all of the other methods. In the case of the wavelet and Gabor methods, the results are acceptable. However, the overall success rates of the co-occurrence matrix and Fourier transform methods in recognizing defects in knitted fabrics are not acceptable.

This paper suggests a novel methodology for building robust information processing systems based on wavelets and artificial neural networks (ANN) to be applied either in decision‐making tasks based on image information or in signal prediction and modeling tasks. The efficiency of such systems is increased when they simultaneously use input information in its original and wavelet transformed form, invoking ANN technology to fuse the two different types of input. A quality control decision‐making system as well as a signal prediction system have been developed to illustrate the validity of our approach. The first one offers a solution to the problem of defect recognition for quality control systems. The second application improves the quality of time series prediction and signal modeling in the domain of NMR. The accuracy obtained shows that the proposed methodology deserves the attention of designers of effective information processing systems.

The current challenge for forecasting smart meters electricity consumption lies in the uncertainty and volatility of load profiles. Moreover, forecasting the electricity consumption for all the meters requires an enormous amount of time. Most papers tend to avoid such complexity by forecasting the electricity consumption at an aggregated level. This paper aims to forecast the electricity consumption for all smart meters at an individual level. This paper, for the first time, takes into account the computational time for training and forecasting the electricity consumption of all the meters.

A novel hybrid autoregressive-statistical equations idea model with the help of clustering and whale optimization algorithm (ARSEI-WOA) is proposed in this paper to forecast the electricity consumption of all the meters with best performance in terms of computational time and prediction accuracy.

The proposed model was tested using realistic Irish smart meters energy data and its performance was compared with nine regression methods including: autoregressive integrated moving average, partial least squares regression, conditional inference tree, M5 rule-based model, k-nearest neighbor, multilayer perceptron, RandomForest, RPART and support vector regression. Results have proved that ARSEI-WOA is an efficient model that is able to achieve an accurate prediction with low computational time.

This paper presents a new hybrid ARSEI model to perform smart meters load forecasting at an individual level instead of an aggregated one. With the help of clustering technique, similar meters are grouped into a few clusters from which reduce the computational time of the training and forecasting process. In addition, WOA improves the prediction accuracy of each meter by finding an optimal factor between the average electricity consumption values of each cluster and the electricity consumption values for each one of its meters.

Aiming at the shortcomings of EEG signals generated by brain's sensorimotor region activated tasks, such as poor performance, low efficiency and weak robustness, this paper proposes an EEG signals classification method based on multi-dimensional fusion features.

First, the improved Morlet wavelet is used to extract the spectrum feature maps from EEG signals. Then, the spatial-frequency features are extracted from the PSD maps by using the three-dimensional convolutional neural networks (3DCNNs) model. Finally, the spatial-frequency features are incorporated to the bidirectional gated recurrent units (Bi-GRUs) models to extract the spatial-frequency-sequential multi-dimensional fusion features for recognition of brain's sensorimotor region activated task.

In the comparative experiments, the data sets of motor imagery (MI)/action observation (AO)/action execution (AE) tasks are selected to test the classification performance and robustness of the proposed algorithm. In addition, the impact of extracted features on the sensorimotor region and the impact on the classification processing are also analyzed by visualization during experiments.

The experimental results show that the proposed algorithm extracts the corresponding brain activation features for different action related tasks, so as to achieve more stable classification performance in dealing with AO/MI/AE tasks, and has the best robustness on EEG signals of different subjects.

The purpose of this paper is to ensure the stable operation of satellites in orbit and to assist ground personnel in continuously monitoring the satellite telemetry data and finding anomalies in advance, which can improve the reliability of satellite operation and prevent catastrophic losses.

This paper proposes a deep auto-encoder (DAE) satellite anomaly advance warning framework for satellite telemetry data. Firstly, this study performs grey correlation analysis, extracts important feature attributes to construct feature vectors and builds the variational auto-encoder with bidirectional long short-term memory generative adversarial network discriminator (VAE/BLGAN). Then, the Mahalanobis distance is used to measure the reconstruction score of input and output. According to the periodic characteristic of satellite operation, a dynamic threshold method based on periodic time window is proposed. Satellite health monitoring and advance warning are achieved using reconstruction scores and dynamic thresholds.

Experiment results indicate DAE methods can probe that satellite telemetry data appear abnormal, trigger a warning before the anomaly occurring and thus allow enough time for troubleshooting. This paper further verifies that the proposed VAE/BLGAN model has stronger data learning ability than other two auto-encoder models and is sensitive to satellite monitoring data.

This paper provides a DAE framework to apply in the field of satellite health monitoring and anomaly advance warning. To the best of the authors’ knowledge, this is the first paper to combine DAE methods with satellite anomaly detection, which can promote the application of artificial intelligence in spacecraft health monitoring.

The purpose of this paper is to systematically study the research and development history of suspicious transaction reporting (STR) system in China, and introduce the core elements in constructing an intelligent surveillance system which could provide a solution to the situation of low effectiveness and efficiency in Chinese Financial Institutions (FIs) STR procedure nowadays. The solution outputs those falling out of the normal customer behavior profiles instead of only extracting data by the rules issued by authorities.

This paper reviews the latest literature, regulations and guidelines of STR gathered domestically and overseas, and hands out questionnaire surveys to hundreds of software vendors, regulators and FIs, details the current situation of poor deployment of intelligent in China and tells the difficulties of subjective STR decision procedures.

Few Chinese FIs have deployed real intelligent STR systems, most are using rule-based filtering systems conformed to the objective STR supervisory regulations. To change the embarrassing situation, the regulators have tried to introduce self-regulatory mode which allows the FIs to define STR decision procedures themselves. Limited by the FIs’ ability of information sharing and investigation scope, FIs could hardly unveil the whole schema of a money laundering organization. The pursuant objective FIs can reach is to construct a system that could tell what the normal customer behaviors look like and extract all those falling out of the system’s expectations as suspicious activities.

Only the core elements of the total intelligent STR system are discussed, that is, what, why and how about the customer behavior pattern recognition system. Besides this, a total solution should also use a watch list, reporting decision, cases management, risk control, etc.

This paper for the first time argues that the orientation of regulatory rules in China has actually hindered the spreading of really effective intelligent system for these years. The author creatively puts forward a solution to the difficult problem for FIs to spot criminal schema directly, instead the FIs should only be required to determine whether the transactions carrying out currently are falling within the expected behavior pattern scopes, which is under the FIs’ capabilities due to the internationally accepted obligations of “Know Your Customer”.

To foster the grid integration of both electric vehicles (EV) and renewable generators, the purpose of this paper is to investigate the possible synergies between these players so as to jointly improve the production predictability while ensuring a green mobility. It is here achieved by the mean of a grid commitment over the overall power produced by a collaborative system which here gathers a photovoltaic (PV) plant with an EV fleet. The scope of the present contribution is to investigate the conditions to make the most of such an association, mainly regarding to the management strategies and optimal sizing, taking into account forecast errors on PV production.

To evaluate the collaboration added value, several concerns are aggregated into a primary energy criterion: the commitment compliance, the power spillage, the vehicle charging, the user mobility and the battery aging. Variations of these costs are computed over a range of EV fleet size. Moreover, the influence of the charging strategy is specifically investigated throughout the comparison of three managements: a simple rule of thumb, a perfect knowledge deterministic case and a charging strategy computed by stochastic dynamic programming. The latter is based on an original modeling of the production forecast error. This methodology is carried out to assess the collaboration added value for two operators’ points of view: a virtual power plant (VPP) and a balance responsible party (BRP).

From the perspective of a BRP, the added value of PV-EV collaboration for the energy system has been evidenced in any situation even when the charging strategy is very simple. On the other hand, for the case of a VPP operator, the coupling between the optimal sizing and the management strategy is highlighted.

A co-optimization of the sizing and the management of a PV-EV collaborative system is introduced and the influence of the management strategy on the collaboration added value has been investigated. This gave rise to the presentation and implementation of an original modeling tool of the PV production forecast error. Finally, to widen the scope of application, two different business models have been tackled and compared.

This paper aims to focus on investigating the failure mode of fused deposition modeling (FDM) fabricated parts by using acoustic emission (AE) technique.

Considering the special prototyping way of FDM, the failure modes of FDM-fabricated parts were investigated experimentally. One test was carried out and the other two describe what has been observed on this basis. Acoustic emissions are obtained during the tensile process. AE features of peak frequency, energy and amplitude are extracted and preliminarily analyzed. Then, the unsupervised clustering method of k-means is applied to explore the relationship between the failure modes and the AE signals.

Failure modes of filament debonding and breakage can be successfully recognized by the pattern recognition technique of k-means.

The results obtained can help us understand the failure process of FDM printed parts. This will provide an available monitoring method in the application of FDM-fabricated parts.

This paper has investigated and characterized the failure modes of FDM fabricated parts for the first time.