Effective Feature Vector Research Articles

A new adversarial malware detection method based on enhanced lightweight neural network

With the gradual expansion of Android systems from mobile phones to intelligent devices, a huge amount of malware has been found every year. To improve the malware detection performance and reduce its reliance on expert experience, deep learning technology has been widely used. However, as the complexity of deep learning models continues to increase, it rapidly increases the consumption of hardware resources. At the same time, anti-detection technology such as Generative Adversarial Networks (GANs) are widely used to evade Artificial Intelligence (AI)-based detection methods. In this paper, we propose a new classification model based on an improved lightweight neural network that can effectively improve the execution efficiency and detection performance of malware detection methods against adversarial malware samples. First, our method uses local-information-entropy-based image generation technology to construct effective image feature vectors. Then, the performance of the lightweight neural network model ESPNetV2 is improved from four aspects. Finally, a new adversarial malware generation model called Mal-WGANGP is proposed. It can automatically generate a large number of adversarial samples to robust our model. In order to evaluate our method, we construct several experiments and compare the detection performance of our method with 19 other novel efficient neural network detection models. Experimental results show that our image enhancement method and detection model have the highest detection accuracy of adversarial samples.

Multi-feature based extreme learning machine identification model of incipient cable faults

In the operation of medium-voltage distribution cables, the local insulation performance may degrade due to inherent defects, environmental influences, and external forces, leading to consecutive self-recovering latent faults in the cables. If not addressed promptly, these faults may escalate into permanent failures. To address this issue, this paper analyzes the development mechanism and characteristics of latent cable faults. A 10kV low-resistance cable latent fault model based on the Kizilcay arc model is built in the PSCAD/EMTDC platform. Furthermore, the paper analyzes and extracts the time-domain, frequency-domain, and time-frequency domain features of fault current samples. Effective fault feature vectors are constructed using multivariate analysis of variance (MANOVA) and Principal Component Analysis (PCA). Based on the fault feature vectors and Extreme Learning Machine (ELM), an intelligent fault identification model for cable latent faults is developed. The initial parameters of the ELM model are optimized using the Particle Swarm Optimization (PSO) algorithm. Finally, the superiority of the proposed model is validated in terms of classification accuracy, training time, and robustness compared to other machine learning algorithms.

A bearing remaining life prediction method under variable operating conditions based on cross-transformer fusioning segmented data cleaning

Bearing remaining useful life (RUL) prediction research based on deep learning mostly emphasizes model performance and effective feature vectors, overlooking different densities of outlier distributions in vibration signals at varying degradation stages. Moreover, forecasting models focus on capturing cross-time dependencies, ignoring the dependencies between different variables. To solve these problems, this paper proposes an unsupervised segmented data cleaning algorithm and a RUL prediction framework adaptable to variable operating conditions. The method consists of four steps: (1) Multi-domain feature extraction and selection establish a feature vector space reflecting degradation trends. (2) Segmented data cleaning divides degradation stages, using different penalty factors for outlier cleaning. (3) Cleaned vibration signals undergo a second round of multidomain feature engineering and degradation-stage division. (4) A two-stage Cross-Transformer model is used for RUL prediction. The method proposed has been validated on the prognostics and health management (PHM) bearing degradation dataset. In the constant condition prediction task, the root mean square error (RMSE) and mean absolute error (MAE) were improved to 1.88 and 5.78, respectively. In the variable condition prediction task, the proposed method outperformed existing methods, with an improvement of 59.10 % in RMSE, demonstrating strong generalization performance and practical application value.

Fault-Diagnosis Method for Rotating Machinery Based on SVMD Entropy and Machine Learning

Rolling bearings and gears are important components of rotating machinery. Their operating condition affects the operation of the equipment. Fault in the accessory directly leads to equipment downtime or a series of adverse reactions in the system, which brings enormous pecuniary loss to the institution. Hence, it is of great significance to detect the operating status of rolling bearings and gears for fault diagnosis. At present, the vibration method is considered to be the most common method for fault diagnosis, a method that analyzes the equipment by collecting vibration signals. However, rotating-machinery fault diagnosis is challenging due to the need to select effective fault feature vectors, use appropriate machine-learning classification methods, and achieve accurate fault diagnosis. To solve this problem, this paper illustrates a new fault-diagnosis method combining successive variational-mode decomposition (SVMD) entropy values and machine learning. First, the simulation signal and the real fault signal are used to analyze and compare the variational-mode decomposition (VMD) and SVMD methods. The comparison results prove that SVMD can be a useful method for fault diagnosis. Then, these two methods are utilized to extract the energy entropy and fuzzy entropy of the gearbox dataset of Southeast University (SEU), respectively. The feature vector and multiple machine-learning classification models are constructed for failure-mode identification. The experimental-analysis results successfully verify the effectiveness of the combined SVMD entropy and machine-learning approach for rotating-machinery fault diagnosis.

Short‐Term Load Forecasting Based on VMD and Combined Deep Learning Model

Aiming at the nonlinear and non‐stationary characteristics of the power load, in order to improve the accuracy of load forecasting, a short‐term load forecasting method based on variational modal decomposition and the combination of convolutional neural network and bi‐directional long short‐term memory (Bi‐LSTM) network is proposed. The load sequence is decomposed into components with different frequencies by variational modal decomposition, and each component is predicted by Bi‐LSTM. Then the convolution neural network is introduced to process multi‐source data. The load, temperature, day and other data are constructed according to the time sliding window as the input, and the convolution neural network is used to extract the effective feature vector, the eigenvector is constructed in a time series and used as the input data of Bi‐LSTM network. Taking the public data set provided by a US public utility department as an example, the proposed model is compared with other methods. The results show that the proposed method can better track the change of load and effectively improve the accuracy of short‐term load forecasting. © 2023 Institute of Electrical Engineer of Japan and Wiley Periodicals LLC.

Residual-Electrical-Endurance Prediction of AC Contactor Based on CNN-GRU

AC contactors are used frequently in various low-voltage control lines, so remaining-life prediction for them can significantly improve the operational reliability of power control systems. To address the problem that the existing AC contactor remaining-life prediction methods do not make full use of the correlation between previous and later states in the degradation process, a CNN-GRU (convolutional neural network-gated recurrent unit) method for AC-contactor remaining-life prediction is proposed. Firstly, the entire cycle of an AC contactor’s degradation data is obtained through a whole-life test, from which the characteristic parameters that effectively reflect the operating states of the contactor are extracted; secondly, neighborhood component analysis (NCA) and maximal information coefficient (MIC) are used to eliminate the redundant information of multidimensional parameters in order to select the optimal feature subset; and then, CNN is used to compress the feature dimension and mine the regular information between the features, so as to extract the effective feature vectors; finally, taking the AC contactor remaining electrical life as a long time sequence issue, time-series accurate prediction is performed using GRU. It is verified that this model is better than RNN (recurrent neural network), LSTM (long short-term memory) and GRU models in prediction, with an effective accuracy of 96.63%, which effectively supports the feasibility of time-series prediction in the field of the remaining-life prediction of electrical devices.

Support vector machine based disease classification model employing hasten eagle Cuculidae search optimization

SummaryThe enormous growth in the field of medicine utilizes machine learning for the well‐being and robustness of human beings. The emerging rate of the disease due to urbanization or various other global facts increases the mortality rate of the individual every year. The detection and classification of disease at the proper time helps an individual to be alive and healthy. Various methods of machine learning are accomplished to classify the diseases, but there is a necessity for the enhancement in the classification methods. In this research, the disease classification is performed using the proposed hasten eagle Cuculidae search optimization algorithm based support vector machine classifier (HECSO‐SVM) with the hybrid kernel function. The feature extraction with the statistical and correlation features presents the effective feature vector for classification and the curse of dimensionality is handled using the proposed HECSO algorithm. Initially, the data will be collected and preprocessing is performed and the features are extracted and selected using the optimization techniques. The proposed method effectively reduces the computation time and increases the accuracy of classification. The analysis is made based on the four different datasets such as Parkinson, heart, lung, and hepatitis that obtains an accuracy rate of 94.582%, 94.100%, 93.985%, 94.582%.

Defect type identification of thin-walled stainless steel seamless pipe based on eddy current testing

Aimed at eddy current detection of defects in thin-walled stainless steel seamless pipes, an effective detection method for identifying defect types is proposed. First, the empirical mode decomposition (EMD) method is used to process the collected eddy current signals and obtain the principal intrinsic mode function (IMF) components of different defects. The Hilbert-Huang transform (HHT) is used to extract the frequency-domain features of the principal IMF components, which are combined with the time-domain features to form an effective defect feature vector. Then, principal component analysis (PCA) is used to reduce the dimensions of the defect feature vector and the redundant information is removed to obtain the principal component vector of the defect. Finally, two radial basis function (RBF) neural networks are used to identify and classify the defect types and three error evaluation indicators are selected to evaluate the performance of the classification network models.

An Effective Feature Extraction Method of Quantitative Recognition of Ultrasonic Echo for Bonding Quality of Composite Sheet

In this paper, a set of effective ultrasonic echo feature extraction method is proposed for the quantitative identification of thin plate composite bonding quality, which opens up a new research idea for the research of ultrasonic echo feature extraction for quantitative identification of thin plate bonding quality, and also lays a foundation for further quantitative identification of bonding structure debonding degree of thin plate. In addition to the time and frequency domain analysis, the empirical mode time-frequency analysis method is used to highlight the time-frequency characteristics of the echo signal, improve the resolution of the signal analysis, and expand the space for the echo signal feature extraction. The feature validity evaluation theory of fault diagnosis is introduced to evaluate the feature validity of high-dimensional signal parameters, and the invalid signal parameters are eliminated to obtain an effective echo signal feature set with low dimension. Using the exhaustive search strategy of feature selection theory, geometric distance measurement and information measurement evaluation criteria, and through the classification verification of support vector machine classifier, the effective feature vector for quantitative identification of adhesive quality is obtained.

Self-attention Mechanism based Dynamic Fault Diagnosis and Classification for Chemical Processes

A dynamic fault detection and diagnosis technique based on deep encoder-decoder network with self-attention mechanism is proposed in this paper. Although traditional encoder-decoder networks exhibit capability in extracting the temporal dependencies, the architecture encodes the input sequence into a fixed-length internal representation. This limits the performance of these networks, especially when considering relatively long input sequences. The self-attention mechanism is used to weight the local feature vectors and retain the correlation between the local information of the signal and the process operation state, so as to extract the effective feature vectors. The extracted features are then fed into the bidirectional encoder-decoder network. The resulting deep network is not only generalizing the importance of local temporal feature, but it also allows the interpretable feature representation and classification simultaneously. The experiments on the benchmark Tennessee Eastman process show that the proposed model has better diagnostic performance on receiver operating characteristic (ROC) and precise-recall (PR) curves than the classical diagnostic method based on the long-short term memory (LSTM) network with convolution layers.

The mRMR-CNN based influential support decision system approach to classify EEG signals

Signal analysis methods are important to extract significant information from signals. In this study, it is aimed to develop an efficient and a reliable EEG signal analysis system based on mRMR-CNN structure used selected features. AlexNet and VGG16 pre-trained network are used to extract feature from the data. Four different Models (Model 1, Model 2, Model 3, Model 4) is examined. In addition, a feature selection algorithm namely mRMR was applied to create a more effective feature vector. Filtering with mRMR allows the concentration of related features and minimization of irrelevant features. The deep features are obtained from the fc6 and fc7 layers. To get high performance, the mRMR algorithm is applied to obtain efficient features and the proposed Model 4 is created. The selected properties with mRMR gave the best results for fine and weighted k-NN. With Model 4, more successful results obtained 98.78%, 98.56%, respectively for fine and weighted k-NN.

A predictive model for phishing detection

Nowadays, many anti-phishing systems are being developed to identify phishing contents in online communication systems. Despite the availability of myriads anti-phishing systems, phishing continues unabated due to inadequate detection of a zero-day attack, superfluous computational overhead and high false rates. Although Machine Learning approaches have achieved promising accuracy rate, the choice and the performance of the feature vector limit their effective detection. In this work, an enhanced machine learning-based predictive model is proposed to improve the efficiency of anti-phishing schemes. The predictive model consists of Feature Selection Module which is used for the construction of an effective feature vector. These features are extracted from the URL, webpage properties and webpage behaviour using the incremental component-based system to present the resultant feature vector to the predictive model. The proposed system uses Support Vector Machine and Naïve Bayes which have been trained on a 15-dimensional feature set. The experiments were based on datasets consisting of 2541 phishing instances and 2500 benign instances. Using 10-fold cross-validation, the experimental results indicate a remarkable performance with 0.04% False Positive and 99.96% accuracy for both SVM and NB predictive models.

SSRDVis: Interactive visualization for event sequences summarization and rare detection

This paper presents SSRDVis, a visual approach to effectively summarize event sequences and interactively detect rare behaviors. SSRDVis is mainly composed of three components: (1) a sequence embedding module for learning effective feature vectors of sequences, (2) a sequence grouping and summarization module to find representative clusters and patterns in the dataset, (3) a rare detection module to discover and explain the rare cases. The sequences are embedded into vector space via “mixed-ngram2vec,” which is adapted from “word2vec.” Then, unsupervised learning models could be applied to group similar sequences and detect anomalies in the vector space. Furthermore, sequential pattern graphs are built to provide a compact and semantic summarization of sequences. These components work together to present both overall sequential patterns and abnormal behaviors in one visual interface. We have demonstrated the feasibility of our approach by applying it to analyze Web clickstreams. Experimental results have shown that our approach could help identify noticeable patterns from a large number of event sequences, especially for rare behaviors.

Consistency-Preserving deep hashing for fast person re-identification

Numerous methods have been proposed for person re-identification (Re-ID) with promising performances. While most of them neglect the matching efficiency which is crucial in real-world applications. Recently, several hashing based approaches have been developed, which consider the importance of matching speed in large-scale datasets. Despite the considerable efficiency of these traditional and deep learning based hashing methods, the concomitant matching accuracy reduction is unacceptable in practical application. Towards this end, we propose a novel deep hashing framework, namely Consistency-Preserving Deep Hashing (CPDH), aiming to bridge the gap between the effective high-dimensional feature and low-dimensional binary vector by focusing on the consistency preservation of hash code. First, CPDH designs a new hash structure to extract the hash code. Next, a noise consistency cost is proposed to improve robustness of both hash code and high-dimensional feature. Finally, a topology consistency cost is provided to maintain the ordinal relation between the high-dimensional feature space and Hamming space. Comprehensive experimental results on three widely-used benchmark datasets demonstrate the superior performance of proposed method as compared with existing state-of-the-art approaches.

Improved inverse gravity moment term weighting for text classification

Text classification is one of the popular high dimensional classification problems where providing better feature vector representations explicitly improve classification performances. Thus, assigning appropriate weights to features or terms are crucial for obtaining effective feature vector representations. The methods used for weighting terms in text classification are called term weighting schemes. Although there exist some term weighting schemes for text classification, they are not fully effective and researchers still focus on proposing new term weighting schemes. In this study, two novel term weighting schemes namely SQRT_TF-IGMimp and TF-IGMimp derived from standard inverse gravity moment formula are proposed to improve weighting behaviors of existing TF-IGM scheme especially for some extreme cases. The performances of proposed schemes are compared with two standard IGM based schemes and five other state-of-the-art term weighting methods on both unbalanced (Reuters-21578) and balanced (20 Mini Newsgroups and 20 Newsgroups) datasets with KNN, SVM, and NN classifiers. Micro-F1 and macro-F1 are used as success measures. The experiments are conducted with various different feature sizes to examine the effects of the feature size on the success of weighting. The experimental results showed that the proposed SQRT_TF-IGMimp method generally outperformed all schemes including both standard TF-IGM and SQRT_TF-IGM schemes. However, the proposed TF-IGMimp scheme also showed mostly better performance than standard TF-IGM. To demonstrate validity of the proposed weighting scheme having maximum performance, t-test is also used and it can be stated that the performance gains obtained by the proposed SQRT_TF-IGMimp weighting scheme compared to standard SQRT_TF-IGM are statistically significant.

Hierarchical feature coding model for high-resolution satellite scene classification

High discriminative feature representation is key in remote sensing scene classification. Existing mid-level feature methods for solving the classification show poor performance. The reason includes two aspects. First, the discrimination power of the feature generated by the feature coding method is limited. Second, semantic information hidden in the scene images are not utilized. These essentially prevent them from achieving better performance. To solve these issues, we propose a hierarchical feature coding model with two stacked feature encoding layers. Specifically, in the first coding layer, semantic information from convolutional layers of deep models and complementary structure and spectral features are extracted and encoded into bag of visual word (BOVW) histogram features. Then in the second layer, Dirichlet-based Gaussians mixture model Fisher kernel is adopted to transform the BOVW histogram features to the more discriminative and effective feature vectors. Thus, through feeding the output of the first layer into the second layer, the complex feature representation is refined. Finally, the concatenated feature vectors are put into support vector machine classifier for classification. Experiments on two public high-resolution remote sensing scene datasets demonstrate that the performance of our hierarchical coding method is comparable to the previous state-of-the-art methods, including most multifeature fusion methods and convolutional neural network-based methods.

Grinding Chatter Detection and Identification Based on BEMD and LSSVM

Grinding chatter is a self-induced vibration which is unfavorable to precision machining processes. This paper proposes a forecasting method for grinding state identification based on bivarition empirical mode decomposition (BEMD) and least squares support vector machine (LSSVM), which allows the monitoring of grinding chatter over time. BEMD is a promising technique in signal processing research which involves the decomposition of two-dimensional signals into a series of bivarition intrinsic mode functions (BIMFs). BEMD and the extraction criterion of its true BIMFs are investigated by processing a complex-value simulation chatter signal. Then the feature vectors which are employed as an amplification for the chatter premonition are discussed. Furthermore, the methodology is tested and validated by experimental data collected from a CNC guideway grinder KD4020X16 in Hangzhou Hangji Machine Tool Co., Ltd. The results illustrate that the BEMD is a superior method in terms of processing non-stationary and nonlinear signals. Meanwhile, the peak to peak, real-time standard deviation and instantaneous energy are proven to be effective feature vectors which reflect the different grinding states. Finally, a LSSVM model is established for grinding status classification based on feature vectors, giving a prediction accuracy rate of 96%.

Fake fingerprint liveness detection based on micro and macro features

Fingerprint is the most hopeful biometric authentication that can specifically identify a person from their exclusive features. In the proposed approach, a novel software-based classification method is presented to classify between fake and real fingerprint. The intention of the proposed system is to improve the security of biometric identification system. The statistical techniques are good for micro features but not well for macro. In this paper, we present a novel combination of local Haralick micro texture features with macro features derived from neighbourhood gray-tone difference matrix (NGTDM) to generate an effective feature vector. Combined extracted features of training and testing images are passed to support vector machine for discriminating live and fake fingerprints. The proposed approach is experimented and validated on ATVS dataset and LivDet2011 dataset. The proposed approach has achieved good accuracy and less error rate in comparison with previously studied techniques.

Fake fingerprint liveness detection based on micro and macro features

Fingerprint is the most hopeful biometric authentication that can specifically identify a person from their exclusive features. In the proposed approach, a novel software-based classification method is presented to classify between fake and real fingerprint. The intention of the proposed system is to improve the security of biometric identification system. The statistical techniques are good for micro features but not well for macro. In this paper, we present a novel combination of local Haralick micro texture features with macro features derived from neighbourhood gray-tone difference matrix (NGTDM) to generate an effective feature vector. Combined extracted features of training and testing images are passed to support vector machine for discriminating live and fake fingerprints. The proposed approach is experimented and validated on ATVS dataset and LivDet2011 dataset. The proposed approach has achieved good accuracy and less error rate in comparison with previously studied techniques.

A novel fault diagnosis method for photovoltaic array based on BP-Adaboost strong classifier

Accurate fault diagnosis of photovoltaic (PV) array is important for effective operation of PV systems. The back propagation neural network (BPNN) based classifier model has wide application in fault diagnosis for PV array. Due to insufficient accuracy obtained by using single BPNN, this paper proposes a novel fault diagnosis scheme based on BP-Adaboost strong classifier. Firstly, several indicators constitute an effective feature vector which is applied to build several BPNN based weak classifier models. Secondly, Adaboost algorithm is adopted to build a strong classifier by combining those weak classifiers into the final output with certain weights. Four operation conditions including normal condition, short circuit fault, partial shade fault, open circuit fault can be accurately identified by the proposed method. Dataset from a 1.8 kW grid-connected PV system with 6 × 3 PV array are applied to experimentally test the performance of the developed method.