Search

The promotion of electric vehicles brings notable environmental and economic advantages. Precisely estimating the state of health (SOH) of lithium-ion batteries is crucial for maintaining their efficiency and safety. This study introduces an SOH estimation approach for lithium-ion batteries that integrates multi-feature analysis with a convolutional neural network and kolmogorov-arnold network (CNN-KAN). Initially, we measure the charging time, current, and temperature during the constant voltage phase. These include charging duration, the integral of current over time, the chi-square value of current, and the integral of temperature over time, which are combined to create a comprehensive multi-feature set. The CNN’s robust feature extraction is employed to identify crucial features from raw data, while KAN adeptly models the complex nonlinear interactions between these features and SOH, enabling accurate SOH estimation for lithium batteries. Experiments were carried out at four different charging current rates. The findings indicate that despite significant nonlinear declines in the SOH of lithium batteries, this method consistently provides accurate SOH estimations. The root mean square error (RMSE) is below 1%, with an average coefficient of determination (R2) exceeding 98%. Compared to traditional methods, the proposed method demonstrates significant advantages in handling the nonlinear degradation trends in battery life prediction, enhancing the model’s generalization ability as well as its reliability in practical applications. It holds significant promise for future research in SOH estimation of lithium batteries.

Dynamic ensemble selection has emerged as a promising approach for hyperspectral image classification. However, selecting relevant features and informative samples remains a pressing challenge. To address this issue, we introduce two novel dynamic residual ensemble learning methods. The first proposed method is called multi-features driven dynamic weighted residuals ensemble learning (MF-DWRL). This method leverages various combinations of features to construct classifier pools that incorporate feature differences. The K-Nearest Neighbors algorithm is employed to establish the region of competence (RoC) in the dynamic ensemble selection process. By assessing the performance of the RoC, the feature sets that yield the highest classification accuracy are identified as the optimal feature combinations. Additionally, the classification accuracy is utilized as prior information to guide the residual adjustments of each classifier. The second method, known as features and samples double-driven dynamic weighted residual ensemble learning (FS-DWRL), further enhances the performance of the ensemble. This approach not only considers the selection of feature combinations but also takes into account the informative samples. By jointly optimizing the feature and sample selection processes, FS-DWRL achieves superior classification accuracy compared to existing state-of-the-art methods. To evaluate the effectiveness of the proposed methods, three hyperspectral datasets from China—WHU-Hi-HanChuan, WHU-Hi-LongKou, and WHU-Hi-HongHu—are used for classification experiments. For these datasets, the proposed methods achieve the highest classification accuracies of 90.57 %, 98.77 %, and 91.08 %, respectively. The MF-DWRL and FS-DWRL methods exhibit significant improvements in classification accuracy.

Objective: To address the challenges arising from the rapid growth of text data in the biomedical field, including the problems of irrelevant argument interference and deep semantic association neglect in existing event argument detection methods, as well as the difficulty of multiple event extraction. We aim to propose a new event argument detection method that can accurately mine valuable information from biomedical texts through multi-feature fusion and the question-and-answer paradigm, while also addressing the limitations of existing methods.Methods: We propose an event argument detection method based on multi-feature fusion and the question-answer paradigm. First, we split each event in the sentence into an independent question-and-answer format to reduce the complexity of detection. Then, in order to reduce the interference of irrelevant arguments, we use syntactic distance and external prior knowledge to find the corresponding argument prior template for each event type. Next, we introduce the multi-feature attention mechanism to fully explore the deep semantic features. Finally, we formulate post-processing methods for predefined event structures to form final biomedical events.Results: On the MLEE dataset, our model achieved 62.50% in event extraction of F1 scores, which is superior to other advanced event extraction methods.Conclusion: Our method achieves good performance in the event extraction task and provides strong support for the mining of valuable information in biomedical texts.

The detection and traceability research of massive log information has always been a focal point in Advanced Persistent Threats (APT) studies. Causality relationship technology analyzes attack paths and simplifies graph scale by mapping attack behaviors, but rule-based detection methods often introduce numerous false alarms. To address this issue, this paper proposes a multi-feature provenance graph for the Accurate Alert System (FPGAA). It constructs trace graphs by selecting feature graphs based on the attack types, effectively improving the compression ratio of trace graphs. The FPGAA system utilizes graph neural network technology to efficiently filter out erroneous attack paths generated by false alarm information. We evaluated the system‘s performance using widely recognized classic datasets, including the DARPA TC dataset and the ATLAS dataset, as well as a self-constructed dataset to ensure comprehensiveness and reliability. The results demonstrate that the FPGAA system shows significant improvements in trace graph optimization compared to similar systems, successfully reducing the provenance graph size by over 300 times. Additionally, the FPGAA system effectively overcomes the interference caused by false alarm alerts, successfully filtering out approximately 94% of false alarms. In summary, our system can accurately identify a variety of attack paths, expedite event investigation efficiency through concise contextual alarms, and enhance identification accuracy by increasing features.

Discovering potential drug-target interactions is crucial for advancing pharmacology. In recent years, the development of large-scale DTI datasets has propelled advancements in DTI prediction computational methods. Various deep learning approaches for interaction prediction often rely on sequence data or structural complexity, yet the synergistic integration of diverse bioinformatics and binding site data remains underexploited, constraining prediction precision. Therefore, a novel approach to integrate available data is required to enhance DTI prediction performance. In this paper, we present a novel aggregation prediction model named MDiDTI, designed to facilitate multi-attribute dual interaction learning. The multi-head self-attention interaction network extracts substructure information of drug molecules and pocket information of targets from biomedical data, enabling spatial-level learning of structural attributes. Meanwhile, the dual-weight mapping network aggregates the chemical semantic features of drug-target pairs, facilitating semantic attribute learning at the sequence level. Lastly, the model combines structural and semantic attributes to compute the interaction values for DTI tasks. Performance evaluation metrics were conducted on three mainstream datasets: BioSNAP, BindingDB, and Human. Experimental results indicate that MDiDTI outperforms existing methods and serves as a reliable and highly generalizable tool for DTI prediction.

With the growing number of user-side resources connected to the distribution system, an occasional imbalance between the distribution side and the user side arises, making short-term power load forecasting technology crucial for addressing this issue. To strengthen the capability of load multi-feature extraction and improve the accuracy of electric load forecasting, we have constructed a novel BILSTM-SimAM network model. First, the entirely non-recursive Variational Mode Decomposition (VMD) signal processing technique is applied to decompose the raw data into Intrinsic Mode Functions (IMF) with significant regularity. This effectively reduces noise in the load sequence and preserves high-frequency data features, making the data more suitable for subsequent feature extraction. Second, a convolutional neural network (CNN) mode incorporates Dropout function to prevent model overfitting, this improves recognition accuracy and accelerates convergence. Finally, the model combines a Bidirectional Long Short-Term Memory (BILSTM) network with a simple parameter-free attention mechanism (SimAM). This combination allows for the extraction of multi-feature from the load data while emphasizing the feature information of key historical time points, further enhancing the model's prediction accuracy. The results indicate that the R2 of the BILSTM-SimAM algorithm model reaches 97.8%, surpassing mainstream models such as Transformer, MLP, and Prophet by 2.0%, 2.7%, and 3.6%, respectively. Additionally, the remaining error metrics also show a reduction, confirming the validity and feasibility of the method proposed.

In recent years, significant progress has been made in facial expression recognition methods. However, tasks related to facial expression recognition in real environments still require further research. This paper proposes a tri-cross-attention transformer with a multi-feature fusion network (TriCAFFNet) to improve facial expression recognition performance under challenging conditions. By combining LBP (Local Binary Pattern) features, HOG (Histogram of Oriented Gradients) features, landmark features, and CNN (convolutional neural network) features from facial images, the model is provided with a rich input to improve its ability to discern subtle differences between images. Additionally, tri-cross-attention blocks are designed to facilitate information exchange between different features, enabling mutual guidance among different features to capture salient attention. Extensive experiments on several widely used datasets show that our TriCAFFNet achieves the SOTA performance on RAF-DB with 92.17%, AffectNet (7 cls) with 67.40%, and AffectNet (8 cls) with 63.49%, respectively.

Aiming at the problem that existing emotion recognition methods fail to make full use of the information in the time, frequency, and spatial domains in the EEG signals, which leads to the low accuracy of EEG emotion classification, this paper proposes a multi-feature, multi-frequency band-based cross-scale attention convolutional model (CATM). The model is mainly composed of a cross-scale attention module, a frequency–space attention module, a feature transition module, a temporal feature extraction module, and a depth classification module. First, the cross-scale attentional convolution module extracts spatial features at different scales for the preprocessed EEG signals; then, the frequency–space attention module assigns higher weights to important channels and spatial locations; next, the temporal feature extraction module extracts temporal features of the EEG signals; and, finally, the depth classification module categorizes the EEG signals into emotions. We evaluated the proposed method on the DEAP dataset with accuracies of 99.70% and 99.74% in the valence and arousal binary classification experiments, respectively; the accuracy in the valence–arousal four-classification experiment was 97.27%. In addition, considering the application of fewer channels, we also conducted 5-channel experiments, and the binary classification accuracies of valence and arousal were 97.96% and 98.11%, respectively. The valence–arousal four-classification accuracy was 92.86%. The experimental results show that the method proposed in this paper exhibits better results compared to other recent methods, and also achieves better results in few-channel experiments.

In complex maritime scenarios where the grayscale polarity of ships is unknown, existing infrared ship detection methods may struggle to accurately detect ships among significant interference. To address this issue, this paper first proposes an infrared image smoothing method composed of Grayscale Morphological Reconstruction (GMR) and a Relative Total Variation (RTV). Additionally, a detection method considering the grayscale uniformity of ships and integrating shape and spatiotemporal features is established for detecting bright and dark ships in complex maritime scenarios. Initially, the input infrared images undergo opening (closing)-based GMR to preserve dark (bright) blobs with the opposite suppressed, followed by smoothing the image with the relative total variation model to reduce clutter and enhance the contrast of the ship. Subsequently, Maximally Stable Extremal Regions (MSER) are extracted from the smoothed image as candidate targets, and the results from the bright and dark channels are merged. Shape features are then utilized to eliminate clutter interference, yielding single-frame detection results. Finally, leveraging the stability of ships and the fluctuation of clutter, true targets are preserved through a multi-frame matching strategy. Experimental results demonstrate that the proposed method outperforms ITDBE, MRMF, and TFMSER in seven image sequences, achieving accurate and effective detection of both bright and dark polarity ship targets.

Abstract User portrait has become a research hot spot in the field of knowledge graph in recent years and the rationality of tag extraction directly affects the quality of user portrait. However, most of the current tag extraction methods for portraits only consider the methods based on word frequency statistics and semantic clustering. These methods have some drawbacks: they cannot effectively discover the preferred themes of the enterprise, dynamically update the portrait tags, and adapt to the needs of the enterprise. In this paper, we propose an enterprise adaptive tag extraction method based on multi-feature dynamic portrait (ATEMDP). ATEMDP first uses K-means to measure the similarity between enterprise texts in preference division, and converts similar enterprise text clustering problems into tag feature clusters to obtain the point cluster structure containing the distribution of tag preference topics. In addition, in the multi-feature selection, the professional domain thesaurus is introduced for feature expansion, and the topic text is introduced into the Bert model as a sample set to discover the potential features of the enterprise text. In the end, in dynamic tag extraction, BiLSTM and CNN are used to extract features, and dynamic preference tags are obtained by updating enterprise text. THUCNews data set and Ente-pku data set are used for simulation, and seven other methods are considered in comparison. The experimental results indicate that ATEMDP is not only superior to other conventional methods in accuracy and F1-score, but also effectively solves the dynamic tagging problem of enterprise portrait.

Hyperspectral image classification (HSIC) has garnered increasing attention among researchers. While classical networks like convolution neural networks (CNNs) have achieved satisfactory results with the advent of deep learning, they are confined to processing local information. Vision transformers, despite being effective at establishing long-distance dependencies, face challenges in extracting high-representation features for high-dimensional images. In this paper, we present the multiscale efficient attention with enhanced feature transformer (MEA-EFFormer), which is designed for the efficient extraction of spectral–spatial features, leading to effective classification. MEA-EFFormer employs a multiscale efficient attention feature extraction module to initially extract 3D convolution features and applies effective channel attention to refine spectral information. Following this, 2D convolution features are extracted and integrated with local binary pattern (LBP) spatial information to augment their representation. Then, the processed features are fed into a spectral–spatial enhancement attention (SSEA) module that facilitates interactive enhancement of spectral–spatial information across the three dimensions. Finally, these features undergo classification through a transformer encoder. We evaluate MEA-EFFormer against several state-of-the-art methods on three datasets and demonstrate its outstanding HSIC performance.

Dynamic networks integrates time attributes on the basis of static networks, and it contains multiple connotations such as the complexity and dynamics of the network structure.It is a better thinking object for studying complex network link prediction problems in the real world.Its high application value has attracted much attention in recent years.However, most of the research objects of traditional methods are still limited to static networks, and there are problems such as insufficient utilization of network time-domain evolution information and high time complexity.Combining sociological theory, a novel temporal link prediction method was proposed based on community multi-feature fusion embedding representation.The core idea of this method was to analyze the dynamic evolution characteristics of the network, learn the embedded representation vector of nodes within the community, and effectively fuse multiple features to measure the generation probability of the connection between nodes.The network was divided into several subgraphs by using community detection with collective influence weights and the Similarity index was proposed based on the collective influence.Then, the biased random walk and the Skip-gram were used to get the embedded vectors for every node and the Similarity index was proposed based on the random walk within the community.Integrating the collective influence, multiple central features of the community, and the representation vector learned within the community, the Similarity index was proposed based on the multi-features fusion.Compared with classical temporal link prediction methods, including moving average methods, embedded representation methods, and graph neural network methods, experimental results on six real data sets show that the proposed methods based on the random walk within the community and the multi-features fusion both achieve better prediction performance under the evaluation criteria of AUC.

Dynamic networks integrates time attributes on the basis of static networks, and it contains multiple connotations such as the complexity and dynamics of the network structure.It is a better thinking object for studying complex network link prediction problems in the real world.Its high application value has attracted much attention in recent years.However, most of the research objects of traditional methods are still limited to static networks, and there are problems such as insufficient utilization of network time-domain evolution information and high time complexity.Combining sociological theory, a novel temporal link prediction method was proposed based on community multi-feature fusion embedding representation.The core idea of this method was to analyze the dynamic evolution characteristics of the network, learn the embedded representation vector of nodes within the community, and effectively fuse multiple features to measure the generation probability of the connection between nodes.The network was divided into several subgraphs by using community detection with collective influence weights and the Similarity index was proposed based on the collective influence.Then, the biased random walk and the Skip-gram were used to get the embedded vectors for every node and the Similarity index was proposed based on the random walk within the community.Integrating the collective influence, multiple central features of the community, and the representation vector learned within the community, the Similarity index was proposed based on the multi-features fusion.Compared with classical temporal link prediction methods, including moving average methods, embedded representation methods, and graph neural network methods, experimental results on six real data sets show that the proposed methods based on the random walk within the community and the multi-features fusion both achieve better prediction performance under the evaluation criteria of AUC.

Abstract Unpaired image-to-image translation for the generation field has made much progress recently. However, these methods suffer from mode collapse because of the overfitting of the discriminator. To this end, we propose a straightforward method to construct a contrastive loss using the feature information of the discriminator output layer, which is named multi-feature contrastive learning (MCL). Our proposed method enhances the performance of the discriminator and solves the problem of model collapse by further leveraging contrastive learning. We perform extensive experiments on several open challenge datasets. Our method achieves state-of-the-art results compared with current methods. Finally, a series of ablation studies proved that our approach has better stability. In addition, our proposed method is also practical for single image translation tasks. Code is available at https://github.com/gouayao/MCL.

Webshell is the most common malicious backdoor program for persistent control of Web application systems, which poses a huge threat to the safe operation of Web servers.For most Webshell detection method based on the request packet data for training, the method for web-based Webshell recognition effect is poorer, and the model of training efficiency is low.In response to the above problems, a Webshell malicious traffic detection method based on multi-feature fusion was proposed.The method was characterized by the three dimensions of Webshell packet meta information, packet payload content and traffic access behavior.Combining domain knowledge, feature extraction of request and response packets in the data stream.Transformed into feature extraction information for information fusion, forming a discriminant model that could detect different types of attacks.Compared with the previous research method, the accuracy rate of the method here in the two classification of normal and malicious traffic has been improved to 99.25%.The training efficiency and detection efficiency have also been significantly improved, and the training time and detection time have been reduced by 95.73% and 86.14%.

Webshell is the most common malicious backdoor program for persistent control of Web application systems, which poses a huge threat to the safe operation of Web servers. For most Webshell detection method based on the request packet data for training, the method for web-based Webshell recognition effect is poorer, and the model of training efficiency is low. In response to the above problems, a Webshell malicious traffic detection method based on multi-feature fusion was proposed. The method was characterized by the three dimensions of Webshell packet meta information, packet payload content and traffic access behavior. Combining domain knowledge, feature extraction of request and response packets in the data stream. Transformed into feature extraction information for information fusion, forming a discriminant model that could detect different types of attacks. Compared with the previous research method, the accuracy rate of the method here in the two classification of normal and malicious traffic has been improved to 99.25%. The training efficiency and detection efficiency have also been significantly improved, and the training time and detection time have been reduced by 95.73% and 86.14%.

Sea ice is a significant factor in influencing environmental change on Earth. Monitoring sea ice is of major importance, and one of the main objectives of this monitoring is sea ice classification. Currently, synthetic aperture radar (SAR) data are primarily used for sea ice classification, with a single polarization band or simple combinations of polarization bands being common choices. While much of the current research has focused on optimizing network structures to achieve high classification accuracy, which requires substantial training resources, we aim to extract more information from the SAR data for classification. Therefore we propose a multi-featured SAR sea ice classification method that combines polarization features calculated by polarization decomposition and spectrogram features calculated by joint time-frequency analysis (JTFA). We built a convolutional neural network (CNN) structure for learning the multi-features of sea ice, which combines spatial features and physical properties, including polarization and spectrogram features of sea ice. In this paper, we utilized ALOS PALSAR SLC data with HH, HV, VH, and VV, four types of polarization for the multi-featured sea ice classification method. We divided the sea ice into new ice (NI), first-year ice (FI), old ice (OI), deformed ice (DI), and open water (OW). Then, the accuracy calculation by confusion matrix and comparative analysis were carried out. Our experimental results demonstrate that the multi-feature method proposed in this paper can achieve high accuracy with a smaller data volume and computational effort. In the four scenes selected for validation, the overall accuracy could reach 95%, 91%, 96%, and 95%, respectively, which represents a significant improvement compared to the single-feature sea ice classification method.

Safety issues are one of the main limitations for further application of lithium-ion batteries, and battery degradation is an important causative factor. However, current state-of-health (SOH) estimation methods are mostly developed for a single feature and a single operating condition as well as a single battery material system, which consequently makes it difficult to guarantee robustness and generalization. This paper proposes a data-driven and multi-feature collaborative SOH estimation method based on equal voltage interval discharge time, incremental capacity (IC) and differential thermal voltammetry (DTV) analysis for feature extraction. The deep learning model is constructed based on bi-directional long short-term memory (Bi-LSTM) with the addition of attention mechanism (AM) to focus on the important parts of the features. The proposed method is validated based on a NASA dataset and Oxford University dataset, and the results show that the proposed method has high accuracy and strong robustness. The estimated root mean squared error (RMSE) are below 0.7% and 0.3%, respectively. Compared to single features, the collaboration between multiple features and AM resulted in a 25% error improvement, and the capacity rebound is well captured. The proposed method has the potential to be applied online in an end-cloud collaboration system.

This paper focuses on how to represent the testing face images for multi-feature face recognition. The choice of feature is critical for face recognition. The different features of the sample contribute differently to face recognition. The joint similar and specific learning (JSSL) has been effectively applied in multi-feature face recognition. In the JSSL, although the representation coefficient is divided into the similar coefficient and the specific coefficient, there is the disadvantage that the training images cannot represent the testing images well, because there are probable expressions, illuminations and disguises in the testing images. We think that the intra-class variations of one person can be linearly represented by those of other people. In order to solve well the disadvantage of JSSL, in the paper, we extend JSSL and propose the extended joint similar and specific learning (EJSSL) for multi-feature face recognition. EJSSL constructs the intra-class variant dictionary to represent the probable variation between the training images and the testing images. EJSSL uses the training images and the intra-class variant dictionary to effectively represent the testing images. The proposed EJSSL method is perfectly experimented on some available face databases, and its performance is superior to many current face recognition methods.

To investigate an effective representation model for 3D facial expression recognition, this paper proposes a multi-feature tensor neighborhood preserving embedding (MFTNPE) method, which seeks various attribute features from raw textured shape scan models of 3D facial expression to construct a multi-feature tensor model. With MFTNPE method, we can project the model into a tensor subspace, explore the interactions and correlations among multimodal data and then extract discriminative low-dimensional tensor features for 3D facial expression recognition. In the study, we first introduce a tensor model based on multi-feature integration for 3D facial expression recognition. Via exploring the textural and geometric features of 3D facial expression data, an original multi-feature tensor model is constructed by stacking various attribute maps orderly. Second, we utilize the tensor completion approach as the scheme of data preprocessing to correct the original multi-feature tensor model. Due to the presence of imperfect data in the projected attribute maps, the tensor completion can predict the missing data in tensor by the trace norm of multi-feature tensor model along each mode. Third, we redefine a weighted matrix for use in the orthogonal-based tensor neighborhood preserving embedding algorithm. In the weighted matrix, the class information is used to guarantee the correct neighborhood relationships for multi-feature tensor samples. This effectively avoids the disturbances and confusions between similar samples while preserving spatial structure and inter-class information. Finally, we conduct extensive experiments on two 3D facial expression recognition datasets. The experimental results show that the proposed method has much better efficiency and performance compared with the state-of-the-art methods.

Given the similarity of handwritten formula symbols and various handwriting styles, this paper proposes a squeeze-extracted multi-feature convolution neural network (SE-MCNN) to improve the recognition rate of handwritten formula symbols. The system proposed in this paper integrates the eight-directional feature of the original sequence in the convolutional layer, which significantly compensates for the lost dynamic trajectory information in the handwritten formula symbol. Meanwhile, the joint loss is constructed to improve the discriminability of features in the way of supervised learning, which enlarges the inter-class difference and decreases inner-class similarity. The standard mathematical formula symbol library provided by the Competition Organization on Recognition of Online Handwritten Mathematical Expression (CROHME) is used to verify the effectiveness of the proposed algorithm. Experiments show that the proposed SE-MCNN approach outperforms the state-of-the-art methods even at the condition of without using the data augmentation.

As the Software Define Network (SDN) adopts centralized control logic, it is vulnerable to various types of Distributed Denial of Service (DDoS) attacks. At present, almost all the research work focuses on high-rate DDoS attack against the SDN control layer. Moreover, most of the existing detection methods are effective for high-rate DDoS attack detection of the control layer, while a low-rate DDoS attack against the SDN data layer is highly concealed, and the detection accuracy against this kind of attack is low. In order to improve the detection accuracy of the low-rate DDoS attack against the SDN data layer, this paper studies the mechanism of such attacks, and then proposes a multi-feature DDoS attack detection method based on Factorization Machine (FM). The features extracted from the flow rules are used to detect low-rate DDoS attacks, and the detection of low-rate DDoS attacks based on FM machine learning algorithms is implemented. The experimental results show that the method can effectively detect the low-rate DDoS attack against the SDN data layer, and the detection accuracy reaches 95.80 percent. Because FM algorithm can achieve fine-grained detection for low-rate DDoS attack, which provides a reliable condition for defending against such attacks. Finally, this paper proposes a defense method based on dynamic deletion of flow rules, and carries out experimental simulation and analysis to prove the effectiveness of the defense method, and the success rate of forwarding normal packets reached 97.85 percent.

In the last 70 years, the automatic text summarization work has become more and more important because the amount of data on the Internet is increasing so fast, and automatic text summarization work can extract useful information and knowledge what user's need that could be easily handled by humans and used for many purposes. Especially in people's daily life, news text is the type of text most people are exposed to. In this study, a new automatic summarzation model for news text which based on fuzzy logic rules, multi-feature and Genetic algorithm (GA) is introduced. Firstly, the most important feature is word features, we score each word and extracted words that exceeded the preset score as keywords and because news text is a special kind of text, it contains many specific elements, such as time, place and characters, so sometimes these special news elements can be extracted directly as keywords. Second is sentence features, a linear combination of these features shows the importance of each sentence and each feature is weighted by Genetic algorithm. At last, we use fuzzy logic system to calculate the final score in order to get automatic summarization. The results of the proposed method was compared with other methods including Msword, System19, System21, System 31, SDS-NNGA, GCD, SOM and Ranking SVM by using ROUGE assessment method on DUC2002 dataset show that proposed method outperforms the aforementioned methods.

We propose a forecasting technique based on multi-feature data fusion to enhance the accuracy of an electric vehicle (EV) charging station load forecasting deep-learning model. The proposed method uses multi-feature inputs based on observations of historical weather (wind speed, temperature, and humidity) data as multiple inputs to a Long Short-Term Memory (LSTM) model to achieve a robust prediction of charging loads. Weather conditions are significant influencers of the behavior of EV drivers and their driving patterns. These behavioral and driving patterns affect the charging patterns of the drivers. Rather than one prediction (step, model, or variables) made by conventional LSTM models, three charging load (energy demand) predictions of EVs were made depending on different multi-feature inputs. Data fusion was used to combine and optimize the different charging load prediction results. The performance of the final implemented model was evaluated by the mean absolute prediction error of the forecast. The implemented model had a prediction error of 3.29%. This prediction error was lower than initial prediction results by the LSTM model. The numerical results indicate an improvement in the performance of the EV load forecast, indicating that the proposed model could be used to optimize and improve EV load forecasts for electric vehicle charging stations to meet the energy requirements of EVs.

Abstract Background Automatic vascular segmentation in X-ray angiographic image sequence is of crucial interest, for instance, for better quantifying coronary arteries in diagnostic and interventional procedures. Methods A novel inter/intra-frame constrained vascular segmentation method is proposed to automatically segment vessels in coronary X-ray angiographic image sequence. First, a morphological filter operator is applied to remove structures undergoing the respiratory motion from the original image sequence. Second, an inter-frame constrained robust principal component analysis (RPCA) is utilized to remove the quasi-static structures from the image sequence. Third, an intra-frame constrained RPCA is employed to smooth the final extracted vascular sequence. Fourth, a multi-feature fusion is designed to improve the vascular contrast and the final vascular segmentation is realized by thresholding-based method. Results Experiments are conducted on 22 clinical X-ray angiographic image sequences. The global and local contrast-to-noise ratio of the proposed method are 6.6344 and 4.2882, respectively. And the precision, sensitivity and F1 value are 0.7378, 0.7960 and 0.7658, respectively. It demonstrates that our method is effective and robust for vascular segmentation from image sequence. Conclusions The proposed method is effective to remove non-vascular structures, reduce motion artefacts and other non-uniform illumination caused noises. Also, the proposed method is online which can just process one image per time without re-optimizing the model.

Doppler-radar-based continuous human motion recognition recently has attracted extensive attention, which is a favorable choice for privacy and personal security. Existing results of continuous human motion recognition (CHMR) using mmWave FMCW Radar are not considered the continuous human motion with the high similarity problem. In this paper, we proposed a new CHMR algorithm with the consideration of the high similarity (HS) problem, called as CHMR-HS, by using the modified Transformer-based learning model. As far as we know, this is the first result in the literature to investigate the continuous HMR with the high similarity. To obtain the clear FMCW radar images, the background and target signals of the detected human are separated through the background denoising and the target extraction algorithms. To investigate the effects of the spectral-temporal multi-features with different dimensions, Doppler, range, and angle signatures are extracted as the 2D features and range-Doppler-time and range-angle-time signatures are extracted as the 3D features. The 2D/3D features are trained into the adjusted Transformer-encoder model to distinguish the difference of the high-similarity human motions. The conventional Transformer-decoder model is also re-designed to be Transformer-sequential-decoder model such that Transformer-sequential-decoder model can successfully recognize the continuous human motions with the high similarity. The experimental results show that the accuracy of our proposed CHMR-HS scheme are 95.2% and 94.5% if using 3D and 2D features, the simulation results also illustrates that our CHMR-HS scheme has advantages over existing CHMR schemes.

State-of-health (SOH) estimation is critical for the battery management system. With autonomous ability and superior nonlinear mapping capability, machine learning is now a hot topic in this field. The training sample set of current machine learning methods based on single learner is often the regional data, which leads to a small range of data acquisition and affects the generalization ability of the model. Regard this issue, the idea of ensemble learning is considered, and by generating differential data samples and synthesizing the output of a series of base learners, a good learning performance can be achieved. Furthermore, gray relational analysis is used for feature correlation analysis. In this paper, NASA battery data sets are used to verify and validate the performance of the proposed method, which indicated an enhanced accuracy of the results based on the ensemble learning method. The proposed battery healthy assessment model based on ensemble learning can be concluded to provide highly accurate and stable SOH predictions.

Image dehazing has been a great challenge in the process of adjusting haze images. In this paper, an effective and accurate dehazing method based on atmospheric scattering model is proposed. Since the dark channel is not applicable to sky areas, single-threshold segmentation combined with quad-tree partition technique is adopted to position and estimate the ambient light A* rapidly and accurately. In order to optimize transmittance, we employ a new convolutional network architecture, multi-feature-based bilinear CNN, which can mitigate the halo effect around the abrupt edges and restrain image noise, and the whole transmittance estimation process can be divided into three main parts: feature extraction, nonlinear mapping, and image reconstruction. A large number of outdoor haze image test results show that our optimized method has better experimental results than the existing methods.

This paper presents a Synthetic Aperture Radar (SAR) image classification algorithm based on multi-feature using Fruit Fly Optimization Algorithm (FOA) and Least Square Support Vector Machine (LS-SVM). First, pixel-based information derived from three elements of coherency matrix, six parameters obtained by $H/\alpha /A$ decomposition and Freeman decomposition techniques, and three polarimetric parameters including the total receive power (SPAN), pedestal height, and Radar Vegetation Index (RVI), as well as region-based information derived from eight texture parameters obtained by Grey Level Co-occurrence Matrix (GLCM) are combined to use as the features of land cover. Second, Kernel Principal Component Analysis (KPCA) is used to reduce the dimensionality of the multi-feature data derived from the integration of the pixel-based and region-based information. Third, LS-SVM is used as the classifier in this study due to its fast solving speed and desirable classification capability. Since the input parameters of LS-SVM significantly affect the classification performance, we employ FOA to obtain the optimized input parameters. Finally, the experiments on two fully polarimetric SAR images of various crops with a limited number of samples are implemented by the proposed method and other commonly used methods, respectively. The results show that the proposed method can attain better classification performances compared with other methods.

Quantifying the information content of remote sensing images is considered to be a fundamental task in quantitative remote sensing. Traditionally, the grayscale entropy designed by Shannon’s information theory cannot capture the spatial structure of images, which has prompted successive proposals of a series of neighborhood-based improvement schemes. However, grayscale or neighborhood-based spatial structure is only a basic feature of the image, and the spatial structure should be divided into the overall structure and the local structure and separately characterized. For this purpose, a multi-feature quantification framework for image information content is proposed. Firstly, the information content of optical remote sensing images is measured based on grayscale, contrast, neighborhood-based topology, and spatial distribution features instead of simple grayscale or spatial structure. Secondly, the entropy metrics of the different features are designed to quantify the uncertainty of images in terms of both pixel and spatial structure. Finally, a weighted model is used to calculate the comprehensive information content of the image. The experimental results confirm that the proposed method can effectively measure the multi-feature information content, including the overall and local spatial structure. Compared with state-of-the-art entropy models, our approach is the first study to systematically consider the multiple features of image information content based on Shannon entropy. It is comparable to existing models in terms of thermodynamic consistency. This work demonstrates the effectiveness of information theory methods in measuring the information content of optical remote sensing images.

Given video streams, we aim to correctly detect unsegmented signs related to continuous sign language recognition (CSLR). Despite the increase in proposed deep learning methods in this area, most of them mainly focus on using only an RGB feature, either the full-frame image or details of hands and face. The scarcity of information for the CSLR training process heavily constrains the capability to learn multiple features using the video input frames. Moreover, exploiting all frames in a video for the CSLR task could lead to suboptimal performance since each frame contains a different level of information, including main features in the inferencing of noise. Therefore, we propose novel spatio-temporal continuous sign language recognition using the attentive multi-feature network to enhance CSLR by providing extra keypoint features. In addition, we exploit the attention layer in the spatial and temporal modules to simultaneously emphasize multiple important features. Experimental results from both CSLR datasets demonstrate that the proposed method achieves superior performance in comparison with current state-of-the-art methods by 0.76 and 20.56 for the WER score on CSL and PHOENIX datasets, respectively.

Tree species classification is crucial for forest resource investigation and management. Remote sensing images can provide monitoring information on the spatial distribution of tree species and multi-feature fusion can improve the classification accuracy of tree species. However, different features will play their own unique role. Therefore, considering various related factors about the growth of tree species such as spectrum information, texture structure, vegetation phenology, and topography environment, we fused multi-feature and multi-temporal Sentinel-2 data, which combines spectral features with three other types of features. We combined different feature-combinations with the random forest method to classify Changbai Mountain tree species. Results indicate that topographic features participate in tree species classification with higher accuracy and more efficiency than phenological features and texture features, and the elevation factor possesses the highest importance through the Mean Decrease in Gini (MDG) method. Finally, we estimated the area of the target tree species and analyzed the spatial distribution characteristics by overlay analysis of the Classification 3 result and topographic features (elevation, slope, and aspect). Our findings emphasize that topographic factors have a great influence on the distribution of forest resources and provide the basis for forest resource investigation.

Named entity recognition of forest diseases plays a key role in knowledge extraction in the field of forestry. The aim of this paper is to propose a named entity recognition method based on multi-feature embedding, a transformer encoder, a bi-gated recurrent unit (BiGRU), and conditional random fields (CRF). According to the characteristics of the forest disease corpus, several features are introduced here to improve the method’s accuracy. In this paper, we analyze the characteristics of forest disease texts; carry out pre-processing, labeling, and extraction of multiple features; and construct forest disease texts. In the input representation layer, the method integrates multi-features, such as characters, radicals, word boundaries, and parts of speech. Then, implicit features (e.g., sentence context features) are captured through the transformer’s encoding layer. The obtained features are transmitted to the BiGRU layer for further deep feature extraction. Finally, the CRF model is used to learn constraints and output the optimal annotation of disease names, damage sites, and drug entities in the forest disease texts. The experimental results on the self-built data set of forest disease texts show that the precision of the proposed method for entity recognition reached more than 93%, indicating that it can effectively solve the task of named entity recognition in forest disease texts.

Islanding detection is one of the conditions necessary for the safe operation of the microgrid. The detection technology should provide the ability to differentiate islanded operations from power grid disturbances effectively. Given that it is difficult to set the fault threshold using the passive detection method, and because the traditional active detection method affects the output power quality, a microgrid islanding detection method based on the Sliding Window Discrete Fourier Transform (SDFT)-Empirical Mode Decomposition (EMD) and Long Short-Term Memory (LSTM) network optimized by an attention mechanism is proposed. In this paper, the inverter output current and voltage at the point of common coupling (PCC) are transformed by the SDFT. The positive sequence, zero sequence, and negative sequence components of voltage and current harmonics are calculated and reconstructed by adopting the symmetrical component method (SCM). Meanwhile, the current and voltage are decomposed into a mono intrinsic mode function (IMF). The symmetric components of voltage, current, and IMFs are used as inputs to the deep learning algorithm. An LSTM with the features extracted to classify islanding and grid disturbance is proposed. By using the attention mechanism to set the weight values of the features of hidden states obtained by the LSTM network, the proportion of important features increases, which improves the classification effect. MATLAB/Simulink simulation results indicate that the proposed method can effectively classify the islanding state under different working conditions with an accuracy level of 98.4% and a loss value of 0.0725 with a maximal detection time of 66.94 ms. It can also reduce the non-detection zone (NDZ) and detection time and has a certain level of noise resistance. Meanwhile, the problem whereby the active method affects the microgrid power quality is avoided without disturbing the current or power of the microgrid.

Recent works have shown that the resting-state brain functional connectivity hypernetwork, where multiple nodes can be connected, are an effective technique for brain disease diagnosis and classification research. The lasso method was used to construct hypernetworks by solving sparse linear regression models in previous research. But, constructing a hypernetwork based on the lasso method simply selects a single variable, in that it lacks the ability to interpret the grouping effect. Considering the group structure problem, the previous study proposed to create a hypernetwork based on the elastic net and the group lasso methods, and the results showed that the former method had the best classification performance. However, the highly correlated variables selected by the elastic net method were not necessarily in the active set in the group. Therefore, we extended our research to address this issue. Herein, we propose a new method that introduces the sparse group lasso method to improve the construction of the hypernetwork by solving the group structure problem of the brain regions. We used the traditional lasso, group lasso method, and sparse group lasso method to construct a hypernetwork in patients with depression and normal subjects. Meanwhile, other clustering coefficients (clustering coefficients based on pairs of nodes) were also introduced to extract features with traditional clustering coefficients. Two types of features with significant differences obtained after feature selection were subjected to multi-kernel learning for feature fusion and classification using each method, respectively. The network topology results revealed differences among the three networks, where hypernetwork using the lasso method was the strictest; the group lasso, most lenient; and the sgLasso method, moderate. The network topology of the sparse group lasso method was similar to that of the group lasso method but different from the lasso method. The classification results show that the sparse group lasso method achieves the best classification accuracy by using multi-kernel learning, which indicates that better classification performance can be achieved when the group structure exists and is properly extended.

Accurate cotton maps are crucial for monitoring cotton growth and precision management. The paper proposed a county-scale cotton mapping method by using random forest (RF) feature selection algorithm and classifier based on selecting multi-features, including spectral, vegetation indices, and texture features. The contribution of texture features to cotton classification accuracy was also explored in addition to spectral features and vegetation index. In addition, the optimal classification time, feature importance, and the best classifier on the cotton extraction accuracy were evaluated. The results showed that the texture feature named the gray level co-occurrence matrix (GLCM) is effective for improving classification accuracy, ranking second in contribution among all studied spectral, VI, and texture features. Among the three classifiers, the RF showed higher accuracy and better stability than support vector machines (SVM) and artificial neural networks (ANN). The average overall accuracy (OA) of the classification combining multiple features was 93.36%, 7.33% higher than the average OA of the single-time spectrum, and 2.05% higher than the average OA of the multi-time spectrum. The classification accuracy after feature selection by RF can still reach 92.12%, showing high accuracy and efficiency. Combining multiple features and random forest methods may be a promising county-scale cotton classification method.

To address the problem of detecting malicious codes in malware and extracting the corresponding evidences in mobile devices, we construct a consortium blockchain framework, which is composed of a detecting consortium chain shared by test members and a public chain shared by users. Specifically, in view of different malware families in Android-based system, we perform feature modeling by utilizing statistical analysis method, so as to extract malware family features, including software package feature, permission and application feature, and function call feature. Moreover, for reducing false-positive rate and improving the detecting ability of malware variants, we design a multi-feature detection method of Android-based system for detecting and classifying malware. In addition, we establish a fact-base of distributed Android malicious codes by blockchain technology. The experimental results show that, compared with the previously published algorithms, the new proposed method can achieve higher detection accuracy in limited time with lower false-positive and false-negative rates.

Wetland vegetation is an important component of wetland ecosystems and plays a crucial role in the ecological functions of wetland environments. Accurate distribution mapping and dynamic change monitoring of vegetation are essential for wetland conservation and restoration. The development of unoccupied aerial vehicles (UAVs) provides an efficient and economic platform for wetland vegetation classification. In this study, we evaluated the feasibility of RGB imagery obtained from the DJI Mavic Pro for wetland vegetation classification at the species level, with a specific application to Honghu, which is listed as a wetland of international importance. A total of ten object-based image analysis (OBIA) scenarios were designed to assess the contribution of five machine learning algorithms to the classification accuracy, including Bayes, K-nearest neighbor (KNN), support vector machine (SVM), decision tree (DT), and random forest (RF), multi-feature combinations and feature selection implemented by the recursive feature elimination algorithm (RFE). The overall accuracy and kappa coefficient were compared to determine the optimal classification method. The main results are as follows: (1) RF showed the best performance among the five machine learning algorithms, with an overall accuracy of 89.76% and kappa coefficient of 0.88 when using 53 features (including spectral features (RGB bands), height information, vegetation indices, texture features, and geometric features) for wetland vegetation classification. (2) The RF model constructed by only spectral features showed poor classification results, with an overall accuracy of 73.66% and kappa coefficient of 0.70. By adding height information, VIs, texture features, and geometric features to construct the RF model layer by layer, the overall accuracy was improved by 8.78%, 3.41%, 2.93%, and 0.98%, respectively, demonstrating the importance of multi-feature combinations. (3) The contribution of different types of features to the RF model was not equal, and the height information was the most important for wetland vegetation classification, followed by the vegetation indices. (4) The RFE algorithm effectively reduced the number of original features from 53 to 36, generating an optimal feature subset for wetland vegetation classification. The RF based on the feature selection result of RFE (RF-RFE) had the best performance in ten scenarios, and provided an overall accuracy of 90.73%, which was 0.97% higher than the RF without feature selection. The results illustrate that the combination of UAV-based RGB imagery and the OBIA approach provides a straightforward, yet powerful, approach for high-precision wetland vegetation classification at the species level, in spite of limited spectral information. Compared with satellite data or UAVs equipped with other types of sensors, UAVs with RGB cameras are more cost efficient and convenient for wetland vegetation monitoring and mapping.

Synthetic aperture radar (SAR) image classification is an important task in remote sensing applications. However, it is challenging due to the speckle embedding in SAR imaging, which significantly degrades the classification performance. To address this issue, a new SAR image classification framework based on multi-feature fusion and adaptive kernel combination is proposed in this paper. Expressing pixel similarity by non-negative logarithmic likelihood difference, the generalized neighborhoods are newly defined. The adaptive kernel combination is designed on them to dynamically explore multi-feature information that is robust to speckle noise. Then, local consistency optimization is further applied to enhance label spatial smoothness during classification. By simultaneously utilizing adaptive kernel combination and local consistency optimization for the first time, the texture feature information, context information within features, generalized spatial information between features, and complementary information among features is fully integrated to ensure accurate and smooth classification. Compared with several state-of-the-art methods on synthetic and real SAR images, the proposed method demonstrates better performance in visual effects and classification quality, as the image edges and details are better preserved according to the experimental results.

Building Change Detection (BCD) is one of the core issues in earth observation and has received extensive attention in recent years. With the rapid development of earth observation technology, the data source of remote sensing change detection is continuously enriched, which provides the possibility to describe the spatial details of the ground objects more finely and to characterize the ground objects with multiple perspectives and levels. However, due to the different physical mechanisms of multi-source remote sensing data, BCD based on heterogeneous data is a challenge. Previous studies mostly focused on the BCD of homogeneous remote sensing data, while the use of multi-source remote sensing data and considering multiple features to conduct 2D and 3D BCD research is sporadic. In this article, we propose a novel and general squeeze-and-excitation W-Net, which is developed from U-Net and SE-Net. Its unique advantage is that it can not only be used for BCD of homogeneous and heterogeneous remote sensing data respectively but also can input both homogeneous and heterogeneous remote sensing data for 2D or 3D BCD by relying on its bidirectional symmetric end-to-end network architecture. Moreover, from a unique perspective, we use image features that are stable in performance and less affected by radiation differences and temporal changes. We innovatively introduced the squeeze-and-excitation module to explicitly model the interdependence between feature channels so that the response between the feature channels is adaptively recalibrated to improve the information mining ability and detection accuracy of the model. As far as we know, this is the first proposed network architecture that can simultaneously use multi-source and multi-feature remote sensing data for 2D and 3D BCD. The experimental results in two 2D data sets and two challenging 3D data sets demonstrate that the promising performances of the squeeze-and-excitation W-Net outperform several traditional and state-of-the-art approaches. Moreover, both visual and quantitative analyses of the experimental results demonstrate competitive performance in the proposed network. This demonstrates that the proposed network and method are practical, physically justified, and have great potential application value in large-scale 2D and 3D BCD and qualitative and quantitative research.

Detection of weeds and crops is the key step for precision spraying using the spraying herbicide robot and precise fertilization for the agriculture machine in the field. On the basis of k-mean clustering image segmentation using color information and connected region analysis, a method combining multi feature fusion and support vector machine (SVM) was proposed to identify and detect the position of corn seedlings and weeds, to reduce the harm of weeds on corn growth, and to achieve accurate fertilization, thereby realizing precise weeding or fertilizing. First, the image dataset for weed and corn seedling classification in the corn seedling stage was established. Second, many different features of corn seedlings and weeds were extracted, and dimensionality was reduced by principal component analysis, including the histogram of oriented gradient feature, rotation invariant local binary pattern (LBP) feature, Hu invariant moment feature, Gabor feature, gray level co-occurrence matrix, and gray level-gradient co-occurrence matrix. Then, the classifier training based on SVM was conducted to obtain the recognition model for corn seedlings and weeds. The comprehensive recognition performance of single feature or different fusion strategies for six features is compared and analyzed, and the optimal feature fusion strategy is obtained. Finally, by utilizing the actual corn seedling field images, the proposed weed and corn seedling detection method effect was tested. LAB color space and K-means clustering were used to achieve image segmentation. Connected component analysis was adopted to remove small objects. The previously trained recognition model was utilized to identify and label each connected region to identify and detect weeds and corn seedlings. The experimental results showed that the fusion feature combination of rotation invariant LBP feature and gray level-gradient co-occurrence matrix based on SVM classifier obtained the highest classification accuracy and accurately detected all kinds of weeds and corn seedlings. It provided information on weed and crop positions to the spraying herbicide robot for accurate spraying or to the precise fertilization machine for accurate fertilizing.

With the enormous amount of multi-source data produced by various sensors and feature extraction approaches, multi-view clustering (MVC) has attracted developing research attention and is widely exploited in data analysis. Most of the existing multi-view clustering methods hold on the assumption that all of the views are complete. However, in many real scenarios, multi-view data are often incomplete for many reasons, e.g., hardware failure or incomplete data collection. In this paper, we propose an adaptive weighted graph fusion incomplete multi-view subspace clustering (AWGF-IMSC) method to solve the incomplete multi-view clustering problem. Firstly, to eliminate the noise existing in the original space, we transform complete original data into latent representations which contribute to better graph construction for each view. Then, we incorporate feature extraction and incomplete graph fusion into a unified framework, whereas two processes can negotiate with each other, serving for graph learning tasks. A sparse regularization is imposed on the complete graph to make it more robust to the view-inconsistency. Besides, the importance of different views is automatically learned, further guiding the construction of the complete graph. An effective iterative algorithm is proposed to solve the resulting optimization problem with convergence. Compared with the existing state-of-the-art methods, the experiment results on several real-world datasets demonstrate the effectiveness and advancement of our proposed method.

In order to solve the problem of how to quickly and accurately obtain crop images during crop growth monitoring, this paper proposes a deep compressed sensing image reconstruction method based on a multi-feature residual network. In this method, the initial reconstructed image obtained by linear mapping is input to a multi-feature residual reconstruction network, and multi-scale convolution is used to autonomously learn different features of the crop image to realize deep reconstruction of the image, and complete the inverse solution of compressed sensing. Compared with traditional image reconstruction methods, the deep learning-based method relaxes the assumptions about the sparsity of the original crop image and converts multiple iterations into deep neural network calculations to obtain higher accuracy. The experimental results show that the compressed sensing image reconstruction method based on the multi-feature residual network proposed in this paper can improve the quality of crop image reconstruction.

Gearbox is an important structure of rotating machinery, and the accurate fault diagnosis of gearboxes is of great significance for ensuring efficient and safe operation of rotating machinery. Aiming at the problem that there is little common compound fault data of gearboxes, and there is a lack of an effective diagnosis method, a gearbox fault simulation experiment platform is set up, and a diagnosis method for the compound fault of gearboxes based on multi-feature and BP-AdaBoost is proposed. Firstly, the vibration signals of six typical states of gearbox are obtained, and the original signals are decomposed by empirical mode decomposition and reconstruct the new signal to achieve the purpose of noise reduction. Then, perform the time domain analysis and wavelet packet analysis on the reconstructed signal, extract three time domain feature parameters with higher sensitivity, and combine them with eight frequency band energy feature parameters obtained by wavelet packet decomposition to form the gearbox state feature vector. Finally, AdaBoost algorithm and BP neural network are used to build the BP-AdaBoost strong classifier model, and feature vectors are input into the model for training and verification. The results show that the proposed method can effectively identify the gearbox failure modes, and has higher accuracy than the traditional fault diagnosis methods, and has certain reference significance and engineering application value.

Exploring functional interactions among various brain regions is helpful for understanding the pathological underpinnings of neurological disorders. Brain networks provide an important representation of those functional interactions, and thus are widely applied in the diagnosis and classification of neurodegenerative diseases. Many mental disorders involve a sharp decline in cognitive ability as a major symptom, which can be caused by abnormal connectivity patterns among several brain regions. However, conventional functional connectivity networks are usually constructed based on pairwise correlations among different brain regions. This approach ignores higher-order relationships, and cannot effectively characterize the high-order interactions of many brain regions working together. Recent neuroscience research suggests that higher-order relationships between brain regions are important for brain network analysis. Hyper-networks have been proposed that can effectively represent the interactions among brain regions. However, this method extracts the local properties of brain regions as features, but ignores the global topology information, which affects the evaluation of network topology and reduces the performance of the classifier. This problem can be compensated by a subgraph feature-based method, but it is not sensitive to change in a single brain region. Considering that both of these feature extraction methods result in the loss of information, we propose a novel machine learning classification method that combines multiple features of a hyper-network based on functional magnetic resonance imaging in Alzheimer's disease. The method combines the brain region features and subgraph features, and then uses a multi-kernel SVM for classification. This retains not only the global topological information, but also the sensitivity to change in a single brain region. To certify the proposed method, 28 normal control subjects and 38 Alzheimer's disease patients were selected to participate in an experiment. The proposed method achieved satisfactory classification accuracy, with an average of 91.60%. The abnormal brain regions included the bilateral precuneus, right parahippocampal gyrus\hippocampus, right posterior cingulate gyrus, and other regions that are known to be important in Alzheimer's disease. Machine learning classification combining multiple features of a hyper-network of functional magnetic resonance imaging data in Alzheimer's disease obtains better classification performance.

Non-intrusive load monitoring (NILM) is a core technology for demand response (DR) and energy conservation services. Traditional NILM methods are rarely combined with practical applications, and most studies aim to disaggregate the whole loads in a household, which leads to low identification accuracy. In this method, the event detection method is used to obtain the switching event sets of all loads, and the power consumption curves of independent unknown electrical appliances in a period are disaggregated by utilizing comprehensive features. A linear discriminant classifier group based on multi-feature global similarity is used for load identification. The uniqueness of our algorithm is that it designs an event detector based on steady-state segmentation and a linear discriminant classifier group based on multi-feature global similarity. The simulation is carried out on an open source data set. The results demonstrate the effectiveness and high accuracy of the multi-feature integrated classification (MFIC) algorithm by using the state-of-the-art NILM methods as benchmarks.

Due to the merits of Lamb wave to Structural Health Monitoring (SHM) of composite, the Lamb wave-based damage detection and identification technology show a potential solution for the insulation condition evaluation of large generator stator. This was performed in order to overcome the problem that it is difficult to effectively identify the stator insulation damage the using single feature of Lamb wave. In this paper, a damage identification method of stator insulation based on Lamb wave multi-feature fusion is presented. Firstly, the different damage features were extracted from time domain, frequency domain, and fractal dimension of lamb wave signals, respectively. The features of Lamb wave signals were extracted by Hilbert transform (HT), power spectral density (PSD), fast Fourier transform (FFT), and wavelet fractal dimension (WFD). Then, a machine learning method based on support vector machine (SVM) was used to fuse and reconstruct the multi-features of Lamb wave and furtherly identify damage type of stator insulation. Finally, the effect of typical stator insulation damage identification is verified by simulation and experiment.

The drawback of pixel-based change detection is that it neglects the spatial correlation with neighboring pixels and has a high commission ratio. In contrast, object-based change detection (OBCD) depends on the accuracy of the segmentation scale, which is of great significance in image analysis. Accordingly, an object-based approach for automatic change detection using multiple classifiers and multi-scale uncertainty analysis (OB-MMUA) in high-resolution (HR) remote sensing images is proposed in this paper. In this algorithm, the gray-level co-occurrence matrix (GLCM), morphological, and Gabor filter texture features are extracted to construct the input data, along with the spectral features, to utilize the respective advantages of the features and to compensate for the insufficient spectral information. In addition, random forest is used to select the features and determine the optimal feature vectors for the change detection. Change vector analysis (CVA) based on uncertainty analysis is then implemented to select the initial training samples. According to the diversity, support vector machine (SVM), k-nearest neighbor (KNN), and extra-trees (ExT) classifiers are then chosen as the base classifiers for Dempster-Shafer (D-S) evidence theory fusion, and unlabeled samples are selected using an active learning method with spatial information. Finally, multi-scale object-based D-S evidence theory fusion and uncertainty analysis is used to classify the difference image. To validate the proposed approach, we conducted experiments using multispectral images collected by the ZY-3 and GF-2 satellites. The experimental results confirmed the effectiveness and superiority of the proposed approach, which integrates the respective advantages of the pixel-based and object-based methods.

Synthetic aperture radar (SAR) is an important means to detect ocean oil spills which cause serious damage to the marine ecosystem. However, the look-alikes, which have a similar behavior to oil slicks in SAR images, will reduce the oil spill detection accuracy. Therefore, a novel oil spill detection method based on multiple features of polarimetric SAR data is proposed to improve the detection accuracy in this paper. In this method, the self-similarity parameter, which is sensitive to the randomness of the scattering target, is introduced to enhance the discrimination ability between oil slicks and look-alikes. The proposed method uses the Random Forest classification combing self-similarity parameter with seven well-known features to improve oil spill detection accuracy. Evaluations and comparisons were conducted with Radarsat-2 and UAVSAR polarimetric SAR datasets, which shows that: (1) the oil spill detection accuracy of the proposed method reaches 92.99% and 82.25% in two datasets, respectively, which is higher than three well-known methods. (2) Compared with other seven polarimetric features, self-similarity parameter has the better oil spill detection capability in the scene with lower wind speed close to 2⁻3 m/s, while, when the wind speed is close to 9⁻12 m/s, it is more suitable for oil spill detection in the downwind scene where the microwave incident direction is similar to the sea surface wind direction and performs well in the scene with incidence angle range from 29.7° to 43.5°.