500 results about "Class imbalance" patented technology

A system and method for machine learning-based data field validation is proposed that utilizes a specific trained machine learning model data architecture that is adapted to be more resilient against training set class imbalance, using a Siamese triplet LSTM network architecture that uses three LSTMs that are trained together and operate in concert. An example non-limiting practical use includes using the Siamese triplet LSTM network architecture to validate whether free-text data fields include a single jurisdiction in an address or multiple jurisdictions in the address.

The invention provides an RSVP type unbalanced electroencephalogram signal classification method based on a self-adaptive channel mixed attention mechanism and decoupling learning, and aims to improve the feature extraction and classification performance of RSVP electroencephalogram signals. Firstly, collected RSVP electroencephalogram signals are preprocessed, including common average reference, band-pass filtering and standardization processing, and a training set and a test set are divided; then, multi-scale time and space features are extracted through a parallel convolution layer and a space convolution layer, and the features of different levels are fused through an interpolation method; the training process is decoupled into representation learning and classifier learning, the representation learning uses a triple loss function based on difficult sample sampling, and the classifier learning uses the optimized Focal Loss to perform feature learning and classifier optimization. A self-adaptive channel mixed attention mechanism is introduced, and the flexibility and robustness of the model are enhanced through dynamic weighting and multi-level feature fusion. According to the method, the problem of class imbalance in the RSVP task is solved.

The invention discloses a strip steel surface defect detection method based on an attention mechanism and an improved UNet network, relates to the field of industrial image processing, and aims to solve the defects of a traditional detection method in defect positioning precision, multi-scale feature fusion and class imbalance. According to the technical scheme, an improved UNet network architecture is adopted, and the method specifically comprises the following steps: in a coding stage, designing a UELAN feature extraction module to improve parameter efficiency; a coordinate attention mechanism (CA) is introduced to enhance the focusing capability on a defect area; and capturing multi-scale high-level semantic features in combination with a cavity space pyramid pooling module (ASPP). In a decoding stage, a convolutional block attention mechanism (CBAM) is integrated, and a channel and spatial feature fusion are dynamically optimized. Besides, a combined loss function is put forward, focus loss and Dice loss are fused, the problem of boundary uncertainty is relieved in combination with a label fuzzification strategy, and category imbalance is coped with through weighted cross entropy. According to the invention, high-precision pixel-level segmentation of strip steel surface defects is realized.

Methods and systems for selecting a deep learning / machine learning model are described. In one embodiment, a plurality of models are optimized by addressing class imbalances among a first subset of records, each of the plurality of models are trained using the first subset of records, each of the plurality of models are implemented to predict a value of respective known target columns in each of a second subset of records, respective success rates for each of the plurality of models at predicting the respective known target columns for each of the second subset of records are determined, a first of the plurality of models having a highest success rate is selected, and the first of the plurality of models is implemented to decide a target column of a received authorization request based on predictor columns in the received authorization request.

The invention discloses an unmanned aerial vehicle fault diagnosis method and system fusing random convolution data enhancement and deep learning, and effectively alleviates the problem of poor model generalization and robustness caused by the class imbalance of an unmanned aerial vehicle flight data set. According to the method, a random convolution data enhancement technology is provided, convolution calculation is performed on an original fault sample, diversified enhanced time sequence samples which maintain original fault features but introduce diversified noise and deformation into local details are generated, and weighted fusion is performed on the enhanced time sequence samples and the original samples, so that a fault sample library is effectively expanded, and the balance of a data set is improved; in order to efficiently capture the time sequence characteristics of the unmanned aerial vehicle fault signals, a fault diagnosis model based on a time convolution network is adopted, and a decision basis is provided for the accuracy of the fault diagnosis model. According to the method, the problem caused by data set category imbalance is effectively solved, the fault diagnosis performance can be improved, and a new thought and method are provided for the unmanned aerial vehicle fault detection method based on data driving.

The invention relates to the technical field of industrial intelligent inspection driven by artificial intelligence, and discloses a multi-modal property management scene risk point detection system based on artificial intelligence, and the system comprises a multi-modal data collection module which is used for synchronously obtaining visible light images, infrared thermal imaging and Internet of Things sensor data in building facilities; the domain self-adaptive defect generation module is based on a decoupling type generator and a StyleGAN2-ADA framework; a multi-modal feature fusion module; an attention enhancement detection module; a lightweight model compression frame; a real-time detection module; and a continuous learning module. High-fidelity defect samples are generated through the decoupling generator and the StyleGAN2-ADA, the problems of sample scarcity and class imbalance are solved, samples are generated through the cyclic generative adversarial network to cover multiple defect types, the acquisition period of the defect samples is shortened, and the requirement for training real samples is reduced.

The invention relates to a photovoltaic fault detection method based on a CNN-LSTM hybrid neural network. The method comprises the following steps: firstly, collecting photovoltaic generating capacity and meteorological data in the same time period according to a group cascade stage; and secondly, preprocessing the acquired data, then constructing a fault detection neural network model, performing feature extraction on the preprocessed data through a convolutional layer by a fault detection module, enhancing feature diversity by using channel rearrangement, processing time sequence data through LSTM, and finally outputting a fault classification result through a full connection layer. In a training process, a focus loss function is used as a loss function of a model to process a class imbalance problem, and a model weight with optimal performance is stored for subsequent fault detection. The method has the advantages of high precision, fast convergence, strong robustness, good generalization ability and the like, and can effectively improve the accuracy of fault detection and classification of the photovoltaic power generation array and realize accurate positioning of the fault string.

The invention discloses a multi-level feature self-attention malicious URL detection method based on post-training, belongs to the field of malicious URL detection, and overcomes the defects of a traditional detection method in domain adaptability and feature capture by constructing a pre-training language model based on post-training and combining multi-level feature extraction and a dynamic attention mechanism. By introducing three self-supervised learning tasks of a mask language model, a noise language model and domain discrimination, the understanding of the model on URL features is enhanced; by extracting multi-level features from a low order to a high order and dynamically adjusting the weight of a feature layer, malicious links can be identified more accurately; experimental results show that under the challenging conditions of small-scale data, class imbalance, attack resistance and the like, the method is remarkably superior to an existing deep learning model and a traditional method, more effective technical support is provided for network security, and meanwhile, the universality, the robustness, the expandability and the data change response capability of the model are evaluated through experiments.

The invention discloses a cascade multi-scale convolution and modal enhancement brain tumor segmentation method based on a Mama framework, and relates to the technical field of brain tumors. According to the method, an MCME-UNet model integrated with an MCMS module is provided, through a hierarchical feature extraction mechanism and a multi-modal feature fusion strategy optimized by an MEM module, the segmentation precision is remarkably improved while the calculation efficiency is kept, and particularly, an EEM module innovatively applies Sobel operator to be connected with residual errors, so that the tumor boundary continuity and the segmentation accuracy are synchronously improved; moreover, a novel training normal form of a focus Tversky loss function is introduced, so that the problem of class imbalance is effectively solved, the sensitivity of the model to a small tumor region is enhanced, a smoother segmentation boundary can be generated, and the stability and accuracy of a segmentation result are greatly improved.

The invention provides a diabetic retinopathy grading method based on a vision-language pre-training model and a ranking perception prompt strategy. According to the method, an eye ground color image and a corresponding DR grading label are converted into image features and text embedding through an image and text dual-coding mode, matching is carried out through the similarity between the image and the text, and accurate DR grading is achieved. In order to solve the problems of class imbalance and natural sequence modeling, a ranking perception prompt module and a similarity matrix smoothing module are introduced, ranking perception prompt ensures that the model can learn and maintain the sequence relation between different DR levels, and an SMS module effectively relieves the influence caused by data imbalance. The method disclosed by the invention can show relatively high accuracy and robustness in various DR grading tasks, especially under the condition of few samples or unbalanced categories. The method not only improves the grading accuracy, but also provides an effective auxiliary tool for clinical diagnosis, and has a wide application prospect.

The invention provides a self-adaptive sampling method for fault diagnosis under data multi-class imbalance and related equipment, and effectively solves the problems of low sample generation quality, high parameter dependence, high calculation complexity, poor adaptability to complex working conditions and the like. The method comprises the steps of coping with different data feature scenes through a parameter adaptive calculation mechanism, then exploring a global optimal solution in multi-classification modeling accuracy model solution optimization by an evolution mechanism through employing a Newton-Raphson optimizer thought, and finally converting an optimal solution set into various fault samples by using a feature recombination mechanism. Therefore, high-quality sample equalization is realized, small sample multi-class imbalance fault diagnosis is carried out by combining MAESTE and multi-class LS-SVM, and the interpretability of the model is improved.

The present invention relates to a method for diagnosing and predicting the course of Alzheimer's disease (AD)-related dementia after feature extraction of brain images by Variational Autoencoders (VAE) and using Artificial intelligence and Machine learning (AI / ML) algorithms, as compared with non-demented control cases (CN), mild cognitive impairment (MCI), and other non-Alzheimer's dementia (non-ADD) cases. Method 1 extracts the reduced dimensional latent feature vector from the sMRI scans / brain images using a VAE and does multiple sampling from the latent distribution to reduce the class imbalance followed by classification using different AI / ML models. Method 2, the extracted feature vectors from VAE are used as input to a mixture of class-Restricted Boltzmann Machines (cl-RBM) for AD, MCI, and CN classification. Method-1 diagnoses AD and distinguishes it from MCI, CN, and non-ADD using MRI scans with or without additional aids. Method-2 further extends to diagnosis and prediction of AD.

The invention relates to the technical field of data classification, in particular to a distributed unshared unbalanced data-oriented classification method, which comprises the following steps of: firstly, learning a converged basic global model based on a federal learning framework; then, an elastic weight solidification method is utilized to enable each local model to further learn local data distribution under the condition that important information in a previous round of global model is not forgotten so as to solve the problems of extreme class imbalance and general class imbalance data on most clients; the problem that an existing distributed unshared data classification method is basically proposed under a federated learning framework, but model disastrous forgetting is caused due to the fact that local data set end class imbalance exists under the framework and class imbalance data generally exists on most clients is not considered is solved. And a global classification model with strong comprehensive performance is inconvenient to cooperatively train by using the data.

The invention belongs to the technical field of network security, and discloses an Internet of Things attack detection method based on deep learning, which eliminates redundant features through feature selection, enhances inter-class difference and intra-class similarity of selected feature subsets, solves the problem of class imbalance in an attack detection data set by utilizing EQLv2, and improves the attack detection efficiency. And meanwhile, constructing a space-time model to extract space-time information of the sample. According to the method, multi-classification performance evaluation is carried out on an Internet of Things attack detection reference data set NSL-KDD and a CIC-IDS-2017, the advantages of the method are verified from five indexes including the accuracy rate, the precision rate, the recall rate, the F1 score and the false alarm rate, and the effectiveness and feasibility of the method are reflected when the method is compared with a plurality of methods.

The invention discloses an irregular small target identification method under a non-high-definition complex background image. A multi-scale feature pyramid is constructed through bidirectional feature fusion, so that the feature expression ability of a small target is enhanced; applying a space-channel attention module to adaptively highlight target features and suppress complex background interference; by introducing a composite loss function including class balance focus loss and enhanced bounding box regression loss, model training is optimized to deal with class imbalance and improve the positioning precision of an irregular target. A self-adaptive multi-scale detection head is adopted, and dynamic feature fusion and scale perception branches are utilized to realize accurate detection of targets with different sizes; according to the method, the problems of low recognition precision and poor adaptability caused by weak features, background interference and irregular shapes of irregular small targets in low-resolution and complex background images are effectively solved, and the monitoring performance in actual applications such as unmanned aerial vehicle aerial photography and remote monitoring is remarkably improved.

The invention discloses a system and a method for monitoring abnormal network traffic of class imbalance, and belongs to the technical field of network space security. The system comprises a flow preprocessing module, a data enhancement module and a monitoring module. The method comprises the following steps: for known class imbalance network traffic data containing normal traffic and partial abnormal traffic, firstly ensuring data quality through a traffic preprocessing module, and extracting features related to anomaly detection; then the data enhancement module carries out undersampling and oversampling on the preprocessed traffic data, and balances the proportion of normal traffic and abnormal traffic in the data set; and finally, the monitoring module clusters the flow data after enhancement processing, evaluates a clustering result, calculates the distance between an unknown flow data sample and the centroid of the cluster to which the unknown flow data sample belongs, and judges whether the flow is abnormal or not. According to the method, the identification capability of the model on the abnormal traffic is enhanced, so that the model can more effectively distinguish the normal traffic from the abnormal traffic.

The invention discloses an equipment anomaly tracing method and system based on a digital twinborn and graph neural network, and belongs to the technical field of industrial intelligent operation and maintenance and fault diagnosis. The invention provides an innovative solution integrating digital twin high-fidelity simulation and a graph structure deep learning algorithm, aiming at the technical bottlenecks that the generalization performance of an existing data driving method is insufficient under the conditions of fault sample scarcity and category imbalance and the traceability accuracy of unknown and composite faults is poor. The method comprises the following steps: constructing a high-fidelity digital twin integrating multi-dimensional physical attributes and a system topology structure; based on a fault mode, influence and harmfulness analysis method system, constructing a fault mode library comprising a plurality of single fault modes and composite fault modes, and generating an enhanced training data set with accurate labels through an automatic fault injection mechanism; training a graph neural network model with a multi-level attention mechanism by using the data set so as to learn a propagation rule of a fault in a complex system topology; and finally deploying the model to carry out abnormity traceability analysis on real-time industrial Internet of Things monitoring data. According to the method, the fault diagnosis generalization ability and the positioning precision under the sample imbalance condition are remarkably improved.

The invention belongs to the technical field of mangrove forest detection and identification, and discloses a mangrove forest vegetation identification method, system and device and a medium, and the method comprises the steps: obtaining an optical image of a to-be-identified region and a corresponding complete polarization decomposition image; performing image fusion on the optical image of the to-be-identified area and the corresponding complete polarization decomposition image to obtain a to-be-identified fusion image; inputting the to-be-identified fusion image into a mangrove forest vegetation identification model for classification prediction to obtain a mangrove forest vegetation identification result; wherein the mangrove forest vegetation identification model comprises an image segmentation module and a sliding overlapping splicing prediction model which are connected in sequence, and the image segmentation module is an AttU-Net network which is constructed based on a lightweight U-Net neural network and is suitable for mangrove forest identification in a high-similarity environment. The method has higher overall performance, better detail capture capability and better robustness against class imbalance in the aspect of mangrove forest vegetation recognition.

The invention discloses an electrocardiogram arrhythmia classification method and system based on a residual shrinkage network, and relates to the technical field of arrhythmia classification. According to the method, efficient electrocardiogram arrhythmia classification is achieved through multi-link cooperation, and a fine preprocessing, data balance strategy and multi-attention mechanism fusion model is designed; preprocessing provides high-quality input through wavelet denoising, precise R-wave detection and the like; the under-sampling-over-sampling mixed strategy is used for solving class imbalance and improving minority class recognition; the ResTCL-Net is fused with CNN, GRU, RCA, TSA and CLA modules, and signal features are mined in multiple dimensions; the optimization training strategy gives consideration to efficiency and stability, and is matched with comprehensive evaluation to guarantee performance. According to the scheme, the classification accuracy and generalization ability are remarkably improved, and abnormal heart beat recognition is enhanced.

Recommender Systems (RS) tend to recommend more popular items instead of the relevant long-tail items. Mitigating such popularity bias is crucial to ensure that less popular but relevant items are recommended. System described herein analyses popularity bias in session-based RS obtained via deep learning (DL) models. DL models trained on historical user-item interactions in session logs (having long-tailed item-click distributions) tend to amplify popularity bias. To understand source of this bias amplification, potential sources of bias at data-generation stage (user-item interactions captured as session logs) and model training stage are considered by the system for recommendation wherein popularity of item has causal effect on user-item interactions via conformity bias, and item ranking from models via biased training process due to class imbalance. While most existing approaches address only one of these effects, a comprehensive causal inference framework is implemented by present disclosure that identifies and mitigates effects at both stages.

Class imbalance in a training dataset may negatively impact the accuracy of a machine-learning model in classifying rare events that are underrepresented in the training dataset. Training datasets comprising time-series data present a unique challenge. Accordingly, resampling techniques for up-sampling and / or down-sampling a training dataset of time series are disclosed. The up-sampling may respect the temporal correlation of time samples in the time series, while generating synthetic time series that mimic the feature values of time series belonging to the minority class. Down-sampling may be used to fine-tune the ratio of time series belonging to the minority class to the time series belonging to the majority class.

The invention discloses an Internet of Vehicles CAN bus intrusion detection method based on noise perception active learning, and belongs to the technical field of Internet of Vehicles safety and machine learning. The invention aims to solve the technical problems of false label noise interference, high manual labeling cost, high attack missing report rate caused by class imbalance and the like. The core of the method is to execute a noise sensing mixed query strategy in an iterative loop: firstly, generating a pseudo tag through clustering and correcting by using an integrated noise detector; secondly, calculating uncertainty scores and noise probabilities of the samples, fusing the uncertainty scores and the noise probabilities to obtain a comprehensive score, and preferentially selecting the samples with high uncertainty and low noise probabilities; and then adaptively selecting a sampling strategy according to the model performance and applying category balance constraint. The query batch is used to iteratively update the model while dynamically adjusting the classification threshold to reduce the missing report rate. According to the method, the influence of pseudo label noise can be effectively suppressed, and the attack detection precision and generalization capability are remarkably improved with extremely low labeling cost.

The invention discloses a single-channel electroencephalogram signal sleep stage classification method, and belongs to the technical field of deep learning, and the method comprises the steps: obtaining a single-channel electroencephalogram signal, carrying out the preprocessing of the single-channel electroencephalogram signal, and generating a signal segment with a fixed time length; performing multi-scale time-frequency feature extraction on the signal segment to generate a primary time-frequency feature, capturing a sleep stage conversion dependency relationship, and performing time sequence enhancement on the primary time-frequency feature to generate an enhanced time sequence feature; performing frequency spectrum statistical feature extraction on the signal segments to generate frequency spectrum statistical features; fusing the enhanced time sequence features and the frequency spectrum statistical features to generate a comprehensive feature vector; according to the comprehensive feature vector, probability distribution of different sleep stages is output through a main classifier, and a binary judgment result of the sleep stage with the minimum sample size is output through an auxiliary classifier. The method can solve the problems of class imbalance, signal complexity and calculation efficiency.

The invention discloses a width unsupervised domain adaptation two-stage rolling bearing fault diagnosis method and system under a class imbalance scene, relates to the field of rolling bearing fault diagnosis, and aims to solve the problem of lack of a method for effectively solving the domain adaptation problem under class imbalance in bearing fault diagnosis in the prior art. Comprising the following steps: step 1, carrying out dimension reduction on data by adopting rolling bearing time domain data; 2, performing feature extraction on the data by utilizing a width learning network model, and performing feature alignment by adopting manifold embedding distribution alignment to obtain a target domain data pseudo tag; 3, performing sample screening based on uncertainty measurement and cosine similarity between samples, and dividing reliable samples and residual samples; introducing a generative adversarial network WGAN-GP containing gradient penalty to perform data enhancement on the screened reliable samples in the target domain to obtain a balanced sample set; and step 4, taking the balanced sample set as a training set, training the width learning network model, and finally inputting residual samples of the target domain for diagnosis.

The invention provides a bearing signal expansion method and system based on a chaotic particle swarm algorithm and a generative adversarial network, and the method comprises the following steps: firstly, carrying out the preprocessing of a collected bearing acceleration signal, and carrying out the denoising through employing the discrete wavelet transform (DWT) in combination with a Bayesian soft threshold method; secondly, determining a proper signal sample length according to a Nyquist sampling theorem, slicing the signal by adopting a sliding window strategy, inputting the sliced data into a chaos particle swarm algorithm for optimization, and searching an optimized vector which is closest to the structural similarity of the sliced data; and finally, inputting slice data of a real sample into a discriminator, and superposing the optimized feature vector with random disturbance to serve as initial input of a generator. In the training process, the parameters of the generator and the discriminator are mutually confronted and updated until a preset training round is reached. And the small sample problem and the class imbalance problem in bearing fault diagnosis are effectively relieved.

The invention provides a ship noise multi-feature classifier data enhancement method and system based on a multi-fine-grained conditional diffusion model. And compressing a waveform to a potential space through VQ-VAE, extracting a ship type / ship name cross semantic vector by using ResNet, and optimizing clustering in combination with a loss function. And a one-dimensional U-Net conditional diffusion model is constructed, unconditional / conditional model output is dynamically weighted and fused, and the weight is adaptively adjusted according to training loss. In the generation stage, a semantic prototype is constructed by using a high-fine-granularity label, parameters are determined by using low / medium-granularity mean value sampling and Bayesian optimization, and fine-granularity controllable waveform generation is realized. After the generated data is converted into multiple features such as MFCC and Lofar, the generated data and original data are combined to train a classifier, and a virtual class strategy relieves class imbalance. Experiments show that the MSE of generated data and real data is reduced, the classification accuracy is improved, the data diversity and the model generalization ability are remarkably enhanced, and the method is suitable for scenes such as underwater target recognition.

The invention discloses an electric arc detection method based on differential increase and structured attention, and the method comprises the steps: firstly carrying out the collection and preprocessing of a current signal, carrying out the differential enhancement according to a sample type, carrying out the strong enhancement of an electric arc sample, improving the generalization capability, and carrying out the weak enhancement of a normal sample, thereby avoiding the overfitting; the problem of class imbalance is relieved, and the model generalization ability is improved; secondly, obtaining six complementary feature representations of time domain waveform, frequency domain frequency spectrum, time frequency analysis, envelope features, statistical distribution and related features from the differentially enhanced current signal through a multi-modal feature extraction method, and fusing to generate a multi-modal image; then, designing a deep learning model integrated with structured attention, carrying out distinguished attention on different feature analysis areas of the multi-modal image, and directionally enhancing arc features; and finally, dynamically quantifying the trained model, reducing the size of the model and reasoning delay, and supporting efficient deployment of various edge computing devices.

The invention discloses a remote sensing image building semantic segmentation method based on HRNet, and the method comprises the steps: building a multi-scale context aggregation system network architecture, enhancing a receptive field, and effectively extracting the multi-level features of an image; a dynamic region aggregation module (DRAM) is embedded in a high-resolution branch, and a global spatial relationship analysis module (GSRAM) is introduced into a low-resolution branch, so that the capturing capability of a fine structure in a high-resolution image and the global context understanding capability of a low-resolution feature are improved; a Point Rend module is adopted to process edge information during building segmentation, and details and edge performance are optimized through point sampling and point rendering. A Poly Loss framework is introduced to design and optimize a loss function, a focus loss function and a polynomial term are combined, the problem of class imbalance is effectively solved, and the segmentation effect is improved. The invention provides an efficient and accurate method for a remote sensing image building segmentation task, and is suitable for the fields of digital city construction, smart city management and the like.

The invention relates to the technical field of food safety, in particular to a food safety risk assessment method for class imbalance data based on small sample learning. The method comprises the following steps: firstly, acquiring a safety limit value interval of the content of a main hazardous substance, analyzing a mapping relationship between grain and oil food safety and the main hazardous substance, proposing a SWARA-CRITIC method to calculate subjective and objective weights, and calculating a comprehensive weight of the main hazardous substance through combining the weights by a game theory; secondly, researching a grain and oil food risk assessment grade division method, and carrying out risk grade objective division through a Gaussian hybrid clustering algorithm; finally, effective feature extraction is conducted on the unbalanced data set through a prototype network algorithm based on the cosine attention mechanism, a new attention mechanism based on the cosine similarity is designed, and compared with a prototype network, the new attention mechanism replaces a traditional zoom dot product attention mechanism. The new mechanism pays more attention to samples with the same characteristics in the support set and the query set, so that a more stable and effective relation is established, and high-precision risk assessment is realized on the premise of not losing key information. The invention provides a novel food safety risk assessment technology based on a small sample learning method.

The invention discloses an anti-fact generation method for processing class imbalance based on a real sample. The method comprises the following steps: preprocessing input data; performing causal feature selection by adopting a causal discovery algorithm; calculating causal feature tendency scores, and performing matching; carrying out anti-fact generation, forming a synthesized minority class set, and integrating the synthesized minority class set with original data to obtain an enhanced data set; and performing data cleaning on the enhanced data set to obtain a balanced data set. According to the method, a data set is effectively balanced by generating a high-quality and close-to-reality anti-fact sample, so that the performance of a downstream classifier on key indexes is remarkably improved; the feature values of the real instances are combined to ensure that the generated samples are located in a reasonable area of data distribution, so that the credibility and availability of the enhanced data are improved; the generated anti-fact sample is located in a boundary region between the majority class and the minority class, the decision region of the minority class is effectively expanded, and the unique post-cleaning avoids the influence of noise accumulation on model training.