Insufficient Feature Extraction Research Articles

Personalized Course Resource Recommendation Algorithm Based on Deep Learning in the Intelligent Question Answering Robot Environment

Aiming at the problems of difficult-to-extract effective information and insufficient feature extraction in the existing intelligent question answering robot environment, a personalized course resource recommendation algorithm based on deep learning is proposed. Firstly, the potential preferences of users are obtained through course-related data. Secondly, the authors use one-hot coding and embedding to convert word vectors into low-dimensional, dense real-valued vectors and input them into the CIN-GRU model. Finally, the attention mechanism is used to improve the attention of some words and the accuracy of personalized course recommendation. The experiment shows that when the recommended list is 35, the precision, recall, and F1 value of the proposed personalized course recommendation method are 0.862, 0.851, and 0.857, respectively, which are higher than those of the comparison method. Therefore, the performance of the proposed method in sustainable personalized course resource recommendation is better than that of the comparison method.

A model of integrating convolution and BiGRU dual-channel mechanism for Chinese medical text classifications.

Recently, a lot of Chinese patients consult treatment plans through social networking platforms, but the Chinese medical text contains rich information, including a large number of medical nomenclatures and symptom descriptions. How to build an intelligence model to automatically classify the text information consulted by patients and recommend the correct department for patients is very important. In order to address the problem of insufficient feature extraction from Chinese medical text and low accuracy, this paper proposes a dual channel Chinese medical text classification model. The model extracts feature of Chinese medical text at different granularity, comprehensively and accurately obtains effective feature information, and finally recommends departments for patients according to text classification. One channel of the model focuses on medical nomenclatures, symptoms and other words related to hospital departments, gives different weights, calculates corresponding feature vectors with convolution kernels of different sizes, and then obtains local text representation. The other channel uses the BiGRU network and attention mechanism to obtain text representation, highlighting the important information of the whole sentence, that is, global text representation. Finally, the model uses full connection layer to combine the representation vectors of the two channels, and uses Softmax classifier for classification. The experimental results show that the accuracy, recall and F1-score of the model are improved by 10.65%, 8.94% and 11.62% respectively compared with the baseline models in average, which proves that our model has better performance and robustness.

Affect and stress detection based on feature fusion of LSTM and 1DCNN

The impact of emotions on health, especially stress, is receiving increasing attention. It is important to provide a non-invasive affect detection system that can be continuously monitored for a long period of time. Multi-sensor fusion strategies can better improve the performance of affect detection models, but there are also problems such as insufficient feature extraction and poor spatiotemporal feature fusion. Therefore, this study proposes a feature-level fusion method based on long short-term memory and one-dimensional convolutional neural network to extract temporal and spatial features of electrocardiogram, electromyogram, electrical activity, temperature, accelerator and response data, respectively, and then fuse them in a summation fashion for affect and stress detection. In particular, we added the tanh activation function before feature fusion, which can improve the model’s performance. We used the wearable affect and stress detection dataset to train the model, which includes three different emotion states (neutral, stress, and amusement) for three-class emotion classification with accuracy and F1-scores of 87.82% and 86.68%, respectively. Due to the importance of stress, we also studied binary classification for stress detection, where neutral and amusement were combined as non-stress, with accuracy and F1-scores of 94.9% and 94.98%, respectively. The performance of the proposed model outperforms other control models and can effectively improve the performance of affect and stress detection, and promote medical care, health care and elderly care.

Improved Text Matching Model Based on BERT

Text matching is a basic and important task in natural language understanding, this paper proposes a new model BBMC for the problem of insufficient feature extraction ability of existing text matching models, which integrates BiLSTM and multi-scale CNN on the basis of BERT. First, the word embedding representation of the text is obtained by the BERT, and then the semantic features of the text are further extracted by the double-layer BiLSTM, followed by the multi-scale CNN model, the key local features are extracted, and finally the linear and SoftMax function are used to classify. Experimental results on the LCQMC dataset show that the BBMC has been improved to a certain extent compared with other methods, and the accuracy on the test set can be best achieved 88.01%.

YOLOv7-RAR for Urban Vehicle Detection.

Aiming at the problems of high missed detection rates of the YOLOv7 algorithm for vehicle detection on urban roads, weak perception of small targets in perspective, and insufficient feature extraction, the YOLOv7-RAR recognition algorithm is proposed. The algorithm is improved from the following three directions based on YOLOv7. Firstly, in view of the insufficient nonlinear feature fusion of the original backbone network, the Res3Unit structure is used to reconstruct the backbone network of YOLOv7 to improve the ability of the network model architecture to obtain more nonlinear features. Secondly, in view of the problem that there are many interference backgrounds in urban roads and that the original network is weak in positioning targets such as vehicles, a plug-and-play hybrid attention mechanism module, ACmix, is added after the SPPCSPC layer of the backbone network to enhance the network's attention to vehicles and reduce the interference of other targets. Finally, aiming at the problem that the receptive field of the original network Narrows, with the deepening of the network model, leads to a high miss rate of small targets, the Gaussian receptive field scheme used in the RFLA (Gaussian-receptive-field-based label assignment) module is used at the connection between the feature fusion area and the detection head to improve the receptive field of the network model for small objects in the image. Combining the three improvement measures, the first letter of the name of each improvement measure is selected, and the improved algorithm is named the YOLOv7-RAR algorithm. Experiments show that on urban roads with crowded vehicles and different weather patterns, the average detection accuracy of the YOLOv7-RAR algorithm reaches 95.1%, which is 2.4% higher than that of the original algorithm; the AP50:90 performance is 12.6% higher than that of the original algorithm. The running speed of the YOLOv7-RAR algorithm reaches 96 FPS, which meets the real-time requirements of vehicle detection; hence, the algorithm can be better applied to vehicle detection.

Feature Pyramid Networks and Long Short-Term Memory for EEG Feature Map-Based Emotion Recognition.

The original EEG data collected are the 1D sequence, which ignores spatial topology information; Feature Pyramid Networks (FPN) is better at small dimension target detection and insufficient feature extraction in the scale transformation than CNN. We propose a method of FPN and Long Short-Term Memory (FPN-LSTM) for EEG feature map-based emotion recognition. According to the spatial arrangement of brain electrodes, the Azimuth Equidistant Projection (AEP) is employed to generate the 2D EEG map, which preserves the spatial topology information; then, the average power, variance power, and standard deviation power of three frequency bands (α, β, and γ) are extracted as the feature data for the EEG feature map. BiCubic interpolation is employed to interpolate the blank pixel among the electrodes; the three frequency bands EEG feature maps are used as the G, R, and B channels to generate EEG feature maps. Then, we put forward the idea of distributing the weight proportion for channels, assign large weight to strong emotion correlation channels (AF3, F3, F7, FC5, and T7), and assign small weight to the others; the proposed FPN-LSTM is used on EEG feature maps for emotion recognition. The experiment results show that the proposed method can achieve Value and Arousal recognition rates of 90.05% and 90.84%, respectively.

Sentiment recognition model of EEG signals combined with one-dimensional convolution and BiBASRU-AT

To address the problems of low accuracy of EEG emotion sentiment and insufficient feature extraction ability of recurrent model, an EEG sentiment recognition model combining one-dimensional convolution and BiBASRU-AT is proposed. The data set is preprocessed in segments to expand the number of samples, and 62 channel local emotional features are extracted from one-dimensional convolution; The built-in self-attention simple recurrent unit is constructed to capture the multi-channel fusion features and the dependence between channels. The soft attention mechanism identifies the key features that have a great impact on the identification of emotional tendencies, and the linear layer outputs the positive, neutral and negative emotion recognition results. The experimental results on the EEG data set(SEED) show that the model achieves an average classification accuracy of 90.24%, which is higher than the excellent deep learning model compared with the experiment. The built-in self-attention simple recurrent unit has stronger feature capture ability, which proves the effectiveness of the model.

Feature extraction algorithm of an irregular small celestial body in a weak light environment.

This study focuses on creating a crater-matching algorithm to improve the matching rate and address the phenomenon of insufficient feature extraction and mismatching of irregular celestial objects and crater edge information on the dim surface of celestial bodies images. These images were captured by the detector's navigation camera. In order to improve the brightness and clarity of the images, the target images were filtered, denoised, and image-enhanced using the bilateral filtering method and improved histogram equalization algorithm, successively. Then, the enhanced image was extracted and matched using the ORB feature point detection algorithm based on scale invariance, and the feature point mismatch was processed by the Hamming distance screening method. The simulation results revealed that the optimization algorithm effectively improved the imaging quality of the target image in dark and weak light environments, increased the number of feature points extracted, reduced the mismatch of effective feature point pairs, and improved the matching rate.

New fusion features convolutional neural network with high generalization ability on rolling bearing fault diagnosis

SummaryThis study proposes a feature fusion one‐dimensional convolutional neural network algorithm model with a portable non‐dimensionality reduction attention mechanism to resolve the limitations pertaining to the small training sample capacity of conventional deep learning models, single feature‐extractor size, insufficient feature extraction of the bearing faults, and low recognition rate of the bearing health status under variable loads and strong noise. This model regarded raw vibration signal as input, used the feature fusion module to extract multiscale time information, and the convolution‐pooling alternating layer adaptively overcome the time‐dependent limitations. An adaptive batch normalization was introduced to reduce the divergence in sample distribution between the source and target domains and enhance the model generalization capabilities. Moreover, the non‐dimensionality reduction attention mechanism is a more efficient channel weight distribution algorithm that can enhance the anti‐noise performance of the model. After embedding the non‐dimensionality reduction attention mechanism, the proposed network could enhance the anti‐noise performance of the model through a special channel‐weight learning method. It combined the Softmax classification layer to construct a feature extraction feature classification dual intelligent fault diagnosis algorithm. The experiments were performed on the rolling bearing fault dataset. The results revealed that the method provided a rather high generalization ability and stable anti‐noise ability. Simultaneously, the introduction of t‐SNE dimensionality reduction visualization revealed the strong feature‐extraction ability of the deep network model for large‐volume samples.

Detecting Ethereum Ponzi Schemes Through Opcode Context Analysis and Oversampling-Based AdaBoost Algorithm

Due to the anonymity of blockchain, frequent security incidents and attacks occur through it, among which the Ponzi scheme smart contract is a classic type of fraud resulting in huge economic losses. Machine learning-based methods are believed to be promising for detecting ethereum Ponzi schemes. However, there are still some flaws in current research, e.g., insufficient feature extraction of Ponzi scheme smart contracts, without considering class imbalance. In addition, there is room for improvement in detection precision. Aiming at the above problems, this paper proposes an ethereum Ponzi scheme detection scheme through opcode context analysis and adaptive boosting (AdaBoost) algorithm. Firstly, this paper uses the n-gram algorithm to extract more comprehensive contract opcode features and combine them with contract account features, which helps to improve the feature extraction effect. Meanwhile, adaptive synthetic sampling (ADASYN) is introduced to deal with class imbalanced data, and integrated with the Adaboost classifier. Finally, this paper uses the improved AdaBoost classifier for the identification of Ponzi scheme contracts. Experimentally, this paper tests our model in real-world smart contracts and compares it with representative methods in the aspect of F1-score and precision. Moreover, this article compares and discusses the state of art methods with our method in four aspects: data acquisition, data preprocessing, feature extraction, and classifier design. Both experiment and discussion validate the effectiveness of our model.

A Fuzzy-Boundary Enhanced Trident Network for Parcel Extraction in the Urban–Rural Area

As the basic unit of farmland, parcel is crucial for remote sensing tasks, such as urban management. Previous studies of farmland parcels extraction are based on boundary detection and instance segmentation methods. However, these methods perform poorly in the parcels with complex shape and fuzzy boundary due to the insufficient feature extraction capability. Moreover, for the lack of multi-scale features extraction and fusion, they are difficult to extract different scale farmland parcels accurately. Based on these issues, we propose a Fuzzy-Boundary Enhanced Trident Network, named FBETNet, to enhance the feature of fuzzy boundary and generate multi-scale parcels. First, a semantic-guided multi-task strategy is introduced in order to enhance the feature of fuzzy boundary. Second, we design a multi-scale trident module to further improve the performance of multi-scale feature extraction. Finally, a adversarial data augmentation strategy is employed in the training phase to strengthen the robustness and stability of out proposed method. Experiments show that our proposed method improves significantly in both accuracy and visualization, especially for the parcels with fuzzy boundary and complex shape.

Patent Text Classification based on Deep Learning and Vocabulary Network

Patent documents are a special long text format, and traditional deep learning methods have insufficient feature extraction ability, which results in a weaker classification effect than ordinary text. Based on this, this paper constructs a text feature extraction method based on the lexical network, according to the inner relation between words and classification. Firstly, the inner relationship between words and classification was obtained from linear and probability dimensions and the lexical network were constructed. Secondly, the lexical network is fused with the features extracted from the deep learning model. Finally, the fusion features are trained in the original model to get the final classification result. T This method is a classification enhancement method that can classify patent text alone or enhance the accuracy of various types of neural networks in patent text classification. Experimental results demonstrate that the accuracy of BERT combined with lexical network method is as high as 82.73%, and the accuracy of lexical network method combined with CNN and LSTM is increased by 2.19% and 2.25% respectively. In addition, it was demonstrated that the lexical network feature extraction method accelerated the convergence speed of the model during training and improved the classification ability of the model in Chinese patent texts.

Facial Expression Recognition Using Multi-Branch Attention Convolutional Neural Network

Facial expression recognition technology was extensively used. This paper develops a multi-branch attention convolutional neural network based on a multiple-branch structure to recognize facial expressions. First, features are extracted from facial images in a multi-branch architecture, and features from three branches are fused. Second, to address the issue of insufficient feature extraction and poor recognition performance, the Convolutional Block Attention Module is added as attention module. Third, our model reduces parameters and computation loads by using depth-wise separable convolutions. The experiments on the FER2013, FERPLUS, and CK+ datasets show that the recognition rates of the proposed model are 69.49%, 84.633%, and 99.39%, respectively. The proposed method has a higher efficiency in extracting image features than traditional deep learning counterparts and achieves high accuracy without complicated artificial feature technology.

Unsupervised multimodal domain adversarial network for time series classification

Unsupervised Domain Adaptation (UDA) is an ideal transfer learning method, which can use labeled source data to improve the classification performance of unlabeled target data. At present, the UDA methods for Time Series Classification (TSC) only use time-domain data or frequency-domain data as the input, and ignore fusing them, resulting in insufficient feature extraction and inaccurate source-target distribution alignment. Therefore, we propose an unsupervised Multimodal Domain Adversarial Network (MDAN) for TSC tasks. Specifically, we adopt two feature extractors for the time-domain and frequency-domain feature representations, and employ-three classifiers to perform TSC of source data for training the two feature extractors; Then, we fuse the time-domain and frequency-domain feature representations of source and target data, respectively, input them into the unified domain discriminator for unsupervised multimodal domain adversarial learning, and combine the proposed Time-Frequency-domain Joint Maximum Mean Discrepancy (TF-JMMD) to accurately align the source-target distributions; Finally, we select CNN or ResNet18 as the feature extractors to carry out comprehensive experiments, and the results demonstrate the SOTA performance of MDAN.

Intelligent diagnosis method of bearing fault based on ICEEMDAN and Ghost-IRCNN

As a critical component of rotating machinery, the health status of bearings is of great significance to ensure the safety of machine operation. Since the vibration signals are interfered by noise during the machine operation, the traditional bearing fault diagnosis model has insufficient feature extraction ability, low diagnostic accuracy, and slow convergence speed. An intelligent bearing fault diagnosis method is proposed based on improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and inverted residual convolution neural network integrated with Ghost module (Ghost-IRCNN). Firstly, ICEEMDAN is used to decompose the signals, and the kurtosis-correlation coefficient-margin factor index is proposed to filter and reconstruct the important modal components. Then, the multi-scale sample entropy is used to construct the feature vectors. Finally, the Ghost module, convolution block attention module, and inverted residual bottleneck are combined to build Ghost-IRCNN to realize bearing fault classification and diagnosis. The experimental results show that the proposed method can extract bearing fault features effectively, and the diagnosis accuracy can reach more than 98.80% under complex conditions.

A Lightweight Bearing Fault Diagnosis Method Based on Multi-Channel Depthwise Separable Convolutional Neural Network

The existing rolling bearing fault diagnosis method based on the deep convolutional neural network has the issues of insufficient feature extraction ability, poor anti-noise ability, and a large number of model parameters. A lightweight bearing fault diagnosis method based on depthwise separable convolutions is proposed. The proposed method can simultaneously extract different features from vibration signals in different directions to enhance the stability of the diagnosis model. The lightweight unit based on depthwise separable convolutions in the feature extraction layer reduces the size of the model and the number of parameters that need to be learned. The vibration signals of bearings in different directions are converted into time-frequency signals by the short-time Fourier transform (STFT) and then into pictures as the input of the model. In order to verify the effectiveness and generalization of the method, this paper uses the gearbox data set of Southeast University and the CWRU (Case Western Reserve University) bearing data set for experiments. Comparisons of bearing fault diagnosis results using the proposed model with other classical deep learning models are implemented. The results show that the proposed model is superior to other classical deep learning models; thus, it has a smaller model size, higher accuracy, and less computation burden. Compared with using a single-direction vibration signal as input, the proposed model that uses multiple vibration signals in different directions as input has more accuracy.

Residual learning-based two-stream network for RGB-T object tracking

RGB-T object tracking is a branch of visual tracking that has been widely applied to many fields, such as intelligent transportation and urban monitoring. Due to the interference of background clutter and occlusion, the existing trackers still suffer from the problems of unreasonable modal fusion strategy, insufficient feature extraction, and loss of semantic information. To solve these problems, we propose a residual learning-based two stream network for RGB-T object tracking. The overall feature extraction network is composed of three branches, and multi-layer convolutions are utilized to extract the features of visible, thermal, and fused images, respectively. First, aiming at improving the effectiveness of feature extraction, a weight generation module for hierarchical feature weight calculation is designed to guide the direction of feature fusion. Then, the residual block is employed to replace the single-layer convolution in order to increase the depth of the network, by which deeper semantic features are learned and the loss of semantic information is alleviated. Finally, a loss function with a penalty term is developed to adjust our network toward the direction of the best tracking performance of the dual modalities. This overcomes the negative impact of poor mode on model training. Experiments implemented on public RGB-T datasets indicate that our algorithm outperforms the recent state-of-the-art trackers in terms of both precision rate and success rate. Compared with the second-best comparison algorithm, our tracker improves the above two metrics by 0.4% and 1.5%, respectively. Our codes are available at https://github.com/MinjieWan/Residual-learning-based-two-stream-network-for-RGB-T-object-tracking.

Research on Improved Traffic Flow Prediction Network Based on CapsNet

Traffic flow prediction is the basis and key to the realization of an intelligent transportation system. The road traffic flow prediction of city-level complex road network can be realized using traffic big data. In the traffic prediction task, the limitation of the convolutional neural network (CNN) for modeling the spatial relationship of the road network and the insufficient feature extraction of the shallow network make it impossible to accurately predict the traffic flow. In order to improve the prediction performance of the model, this paper proposes an improved capsule network (MCapsNet) based on capsule network (CapsNet). First, in the preliminary feature extraction stage, a depthwise separable convolutional block is added to expand the feature channel to enrich channel information. Subsequently, in order to strengthen the reuse of important features and suppress useless information, channel attention is used to selectively reinforce learning of extended channel information so that the network can extract a large number of high-dimensional important features and improve the ability of network feature learning and expression. At the same time, in order to alleviate the feature degradation during training and the channel collapse problem easily caused by deep convolution, a shortcut connection, and a modified linear bottleneck layer structure are added to the convolution layer. The bottleneck layer adds the depth convolution and channel attention connection to the residual block of the network. Finally, the deep local feature information extracted from the improved convolutional layer is vectorized into the form of a capsule, which can more accurately model the details of road network attributes and features and improve the model expression power and prediction performance. The network is tested on the Wenyi Road dataset and the public dataset SZ-taxi. Compared with other models, the evaluation indicators of MCapsNet are better than other models in the tests of different time periods and predictors. Compared with CapsNet, the RMSE index of MCapsNet is reduced by 10.50% in the full period of Wenyi Road, 4.66% in the peak period, 9.78% in the off-peak period, and 6.07% in the SZ-tax dataset. The experimental results verify the effectiveness of the model improvement.

Motor Imaging EEG Signal Recognition of ResNet18 Network Based on Deformable Convolution

EEG signals with a weak amplitude, complex background noise, randomness, significant individual differences, and small data volume lead to insufficient feature extraction and low classification accuracy. Spurred by these concerns, this paper proposes a motor imaging EEG signal classification method based on fusing the improved ResNet18 network with the deformable convolutional network (DCN). Specifically, the original signal’s local airspace characteristics are enhanced by the common spatial patterns (CSP), and the time-frequency domain characteristics are displayed using the short-time Fourier transform (STFT). Then, the signal is converted into a time-frequency map, where a deformable convolution is applied to capture the contour characteristics of the time-frequency map. This strategy solves the problems of traditional convolution related to hard rules, i.e., the convolutional kernel shape can only be a square or rectangular core and cannot be dynamically changed according to the recognition target, resulting in a low recognition rate, prohibiting the network from extracting hidden features and affording enhanced identification and classification. Experimental results demonstrate that our method attains an average classification accuracy on a two-classification and two four-classification motor imaging EEG signals of 90.30%, 86.50%, and 88.08%, respectively, which is much higher than current work, proving our method’s effectiveness.

State-of-Health Prediction Using Transfer Learning and a Multi-Feature Fusion Model.

Existing data-driven technology for prediction of state of health (SOH) has insufficient feature extraction capability and limited application scope. To deal with this challenge, this paper proposes a battery SOH prediction model based on multi-feature fusion. The model is based on a convolutional neural network (CNN) and a long short-term memory network (LSTM). The CNN can learn the cycle features in the battery data, the LSTM can learn the aging features of the battery over time, and regression prediction can be made through the full-connection layer (FC). In addition, for the aging differences caused by different battery operating conditions, this paper introduces transfer learning (TL) to improve the prediction effect. Across cycle data of the same battery under 12 different charging conditions, the fusion model in this paper shows higher prediction accuracy than with either LSTM and CNN in isolation, reducing RMSPE by 0.21% and 0.19%, respectively.