137 results about "Data complexity" patented technology

The invention discloses a multivariable time sequence prediction method based on GRU and a computer program product. The method comprises the following steps: firstly, introducing the dynamic characteristics of self-adaptive different time sequences of self-defined time sequence decomposition, and splitting an original sequence into trend, seasonal and residual components so as to reduce the data complexity and improve the interpretability; afterwards, a value embedding module is used for unifying feature representation so as to ensure that the model fully captures the time dependency relationship; in the modeling stage, the model adopts a multi-channel recurrent neural network to independently model the three types of components so as to reduce the interference between modes and improve the learning ability. In the prediction stage, a feature splicing strategy is adopted, information of each component is integrated, and richer time sequence representation is provided. In addition, a segmented prediction strategy is designed for the model, the prediction process is divided into multiple time periods, prediction information is combined, error accumulation is reduced, and the stability and robustness of long-sequence prediction are improved. Experimental results show that the prediction precision can be improved on different data sets.

The invention discloses a data asset classification method and system based on big data, and the method comprises the steps: carrying out the standardization processing of multi-source data assets in an enterprise business system, constructing a knowledge graph through correlation analysis, dynamically updating the graph according to the flow condition of the data assets, determining key data assets according to the use frequency of the data assets, and carrying out the classification of the key data assets. The method comprises the steps of obtaining a knowledge graph, obtaining an association path and association strength from the knowledge graph, carrying out foreground prediction, clustering prediction results according to data complexity, calculating a data asset evaluation average value of each cluster as a clustering center index, classifying all data assets according to the clustering center index, and forming a data asset portrait. And mutually isolated distributed storage is executed, and a minimum storage principle is followed to optimize life cycle management of data assets. According to the method, the utilization rate of data assets is increased, data security management is enhanced, an enterprise is supported to make a more intelligent data-driven decision, and meanwhile, the method has relatively good interpretability.

The invention discloses an intelligent management system and method for knowledge graph quality evaluation and self-repairing, belongs to the technical field of knowledge graphs, and aims to solve the problems that in traditional knowledge graph management, manual auditing efficiency is low, an effective automatic repairing means is lacked, and data complexity and real-time changes are difficult to deal with. The system firstly collects multi-source heterogeneous data in a target field, cleans the data through a deep learning noise recognition model, extracts entities and relationships by using a natural language processing technology, and adds metadata to convert the entities and relationships into graph structure data; then, a graph framework is defined based on the ontology, entity semantic alignment is achieved in combination with a graph neural network, and a knowledge graph is constructed by complementing implicit relations with the help of a pre-training language model. Then, the quality of the atlas is quantitatively evaluated through a four-layer quality evaluation system, meanwhile, a repair scheme is generated based on vulnerability feature extraction, knowledge base matching and decision fusion, and intelligent self-repair is achieved; the map can be monitored in real time and evaluated regularly, a repair strategy and a knowledge base are optimized through reinforcement learning, it is ensured that the map is kept accurate and time-efficient for a long time, and the practical value is improved.

The invention relates to the technical field of wind speed prediction, in particular to a wind speed prediction method and system based on entropy clustering, and the method comprises the following steps: carrying out the multi-scale decomposition of a wind speed sequence through a permutation entropy algorithm, generating subsequences, carrying out the sliding window probability distribution calculation of a plurality of subsequences, setting a dynamic classification threshold value based on a normalized entropy value, and carrying out the calculation of the sliding window probability distribution. And dividing the sub-sequence set into an entropy classification result set. According to the method, a classification threshold value is dynamically set through sliding window probability distribution, the sensitivity of fixed entropy division to data complexity change is overcome, nonlinear mapping between an original sequence and entropy components is established through GRU training, the feature extraction depth and low component error transmission are enhanced, high, medium and low entropy components are modeled through Bi-RNN, LSTM and SVR, feature extraction with different complexities is matched, and the accuracy of feature extraction is improved. Chaotic mapping optimization fusion weight, multi-target optimization balance component complementarity and dynamic error reference are combined with feedback to adjust a confidence interval, error distribution adaptive tracking is realized, and prediction reliability is enhanced.

The embodiment of the invention discloses a multi-modal large model optimization method and device for a complex task and a medium, and relates to the technical field of multi-modal large models.The method comprises the steps that multi-modal input data are received, modal specific feature extraction is carried out on the multi-modal input data, and initial feature representation of each modal is generated, the multi-modal input data comprises image data, text data and audio data; generating a dynamic parameter adjustment strategy based on a pre-acquired task type and input data complexity, and adjusting calculation parameters of a multi-modal encoder in the multi-modal large model according to the dynamic parameter adjustment strategy to determine a corresponding optimization modal processing sub-module; and performing cross-modal fusion by optimizing the modal processing sub-module and the initial feature expression to generate a fused multi-modal feature expression, and performing semantic analysis and reasoning on the multi-modal feature expression by using a pre-trained multi-modal large model to generate a task output result.

A method for visually identifying crop rows includes acquiring imagery from at least one imaging device operatively connected to an agricultural vehicle while the agricultural vehicle is traversing a field, the imagery includes plants arranged in rows, processing the imagery at a computing device to preserve spatial information along a first dimension while reducing data complexity along a second dimension to generate processed data, analyzing the processed data to identify locations of the rows within the field. The method may include dividing the imagery into multiple sections along a direction of travel of the agricultural vehicle and for each section, performing statistical analysis to generate a one-dimensional intensity profile, applying an edge detection operation to identify predominantly vertical features within the imagery and filtering out predominantly horizontal features to isolate plant stalks, and / or identifying spatial relationships between neighboring rows and using the spatial relationships to validate row identification.

The invention discloses an artificial intelligence large model reasoning acceleration method and system based on a GPU and an NPU, and relates to the technical field of large models. The invention discloses an artificial intelligence large model reasoning acceleration system based on a GPU (Graphics Processing Unit) and an NPU (Network Processing Unit). The system comprises a reasoning information shunting module and a reasoning data acceleration module, according to the method, by combining input data features, model structure dynamic control and heterogeneous computing resource collaborative scheduling, highly-adaptive large model reasoning optimization is realized, accurate matching of computing resources and data complexity is realized, and the overall reasoning throughput and response speed are remarkably improved; a dynamic scheduling mechanism based on a strategy output model is adopted, GPU / NPU calculation tasks are reasonably distributed, resource idling and congestion are avoided, the hardware utilization rate is improved, and the method is particularly suitable for multi-task concurrent and high-frame-rate video scenes.

The invention relates to a ship user behavior self-learning recommendation system based on a large model, and relates to the technical field of ship informatization. According to the system, through collection and fusion of multi-source heterogeneous ship user behavior data, a large-scale pre-training language model (large model) is utilized to carry out deep understanding and semantic mining on massive ship field text information and user behavior sequences, and a ship field knowledge graph or semantic vector space is constructed. The large model can identify and predict potential demands, behavior patterns and preference changes of ship users, and generates highly personalized, accurate and prospective ship service, product, route or information recommendations in combination with real-time operation data and external environment factors. Besides, a user feedback self-learning mechanism is introduced into the system, recommendation strategies and model parameters are continuously optimized according to interaction behaviors and explicit evaluation of the users in modes of reinforcement learning or continuous learning and the like, and intelligent iteration of the system and continuous improvement of the recommendation effect are achieved. According to the method, the challenges of a traditional recommendation system in the aspects of data complexity, semantic gaps and dynamic demand adaptability in the ship field are effectively solved, and the ship operation efficiency and the user satisfaction degree are remarkably improved.

The invention discloses a distributed task processing method and device based on spatio-temporal data complexity and a storage medium, and relates to the technical field of data processing, and the method comprises the steps: extracting a feature vector of spatio-temporal data of a digital twin model, and inputting the feature vector into a trained random forest model; through the trained random forest model, performing classification prediction on the spatio-temporal data according to the feature vector, and outputting a category label corresponding to the feature vector, the category label representing the complexity of the spatio-temporal data; mapping the category labels of the spatio-temporal data into weight coefficients, and formulating distributed task queues based on the weight coefficients, so that the difference of the sum of weights of the task queues of each processing node is minimum to ensure load balance among the distributed task queues; and distributing the distributed task queue to each cluster node for distributed processing. The technical effect of improving the distributed task processing efficiency is achieved.

The invention relates to the technical field of smart park data monitoring, and discloses an artificial intelligence smart park multi-source data monitoring and analysis method. The method comprises the following steps: constructing a park operation state basic model; when the model is identified to be abnormal, data dynamic change characteristics are extracted by utilizing a spatio-temporal data fusion technology, and an abnormal event mode in a specific region is identified by combining deep belief network deep analysis; calculating data complexity by using permutation entropy, optimizing and analyzing an abnormal propagation path in combination with an A star algorithm, and evaluating a diffusion range and an influence path; recognizing a comprehensive risk area, analyzing the environment and equipment state characteristics of the area through a multispectral imaging technology, and evaluating the operation abnormity in combination with real-time multi-source data; and performing accurate intervention on the comprehensive risk area based on the abnormal condition. According to the method, effective integration and analysis of multi-source data of the smart park are realized, abnormity is accurately identified, risks are mastered, a scientific means is provided for park management, and stable and efficient operation of the park is guaranteed.

The invention provides a significant wave height prediction method and device and electronic equipment, and relates to the technical field of sea wave data processing. The method comprises the following steps: acquiring an effective wave height data set for wave height prediction; determining model input data based on the correlation; carrying out two-layer decomposition on the effective wave height; wherein the primary decomposition comprises empirical mode decomposition based on a self-adaptive noise complete set and K-means clustering of mode components; the secondary decomposition comprises the step of optimizing variational mode decomposition of parameters based on a crown porcupine optimization algorithm; inputting model input data into a bidirectional long-short-term memory network based on an attention mechanism for training; the method comprises the steps of firstly screening data with relatively high correlation with the significant wave height as input data to reduce the calculation amount, then performing two-layer decomposition on the significant wave height to reduce the data complexity, and finally fully extracting data features through a bidirectional long-short-term memory network based on an attention mechanism to comprehensively predict the data so as to improve the prediction accuracy of the significant wave height. And the accuracy of predicting the significant wave height is improved.

The invention provides a financial risk prediction method based on a multi-objective ensemble learning algorithm, and aims to solve the problems of multi-risk type comprehensive evaluation, data complexity and nonlinear relation processing, dynamic change adaptability and the like in risk prediction in the financial field. According to the method, a financial time sequence is divided into a plurality of time blocks through a space-time block data preprocessing technology, and a heterogeneous model library is independently trained, so that the adaptability and prediction precision of a model are improved. Meanwhile, an NSGA-II multi-objective optimization algorithm is combined with Gibbs distribution to dynamically distribute weights, the prediction precision and the engineering practicability are balanced, and dynamic optimization of economic indexes and engineering indexes is achieved. In addition, a time decay and mixed attention mechanism is designed, multi-model prediction results are fused, market dynamic changes are further captured, and the accuracy and real-time performance of risk prediction are improved. According to the method, the precision and the real-time response capability of financial risk prediction are effectively improved.

The invention discloses a thermal load prediction method and system based on multi-stage fusion optimization, and the method employs a gray wolf optimization algorithm to replace a conventional empirical method to optimize key parameters of variational mode decomposition, and automatically determines an optimal mode number and a penalty factor through multi-objective optimization. The obtained parameter combination is introduced into a variational mode decomposition algorithm, and then the variational mode decomposition algorithm is utilized to decompose thermal load historical data into a plurality of stable mode components, so that the complexity and noise interference of the data are reduced, and the decomposition quality of variational mode decomposition is remarkably improved; then, the hyper-parameters of the long-short-term memory network are optimized by combining a sparrow search algorithm, so that an optimal hyper-parameter combination is found, and finally, a thermal load data set is predicted through a long-short-term memory model under the obtained optimal hyper-parameter combination.

The invention belongs to the technical field of data security, and particularly relates to a non-intrusive power load identification method and system based on an intelligent fusion terminal, and the method comprises the steps: collecting power load data, and dividing an analysis window based on the power load data; calculating a volatility index corresponding to the analysis window based on the data in the analysis window so as to represent the data complexity in the analysis window; determining the number of iterations for encryption based on the volatility index; and executing a compression algorithm on the data in the analysis window to generate compressed data, and executing multiple rounds of encryption perturbations determined by the number of iterations on the compressed data to generate a dynamically encrypted compressed data stream. According to the method, the privacy protection intensity can be adaptively adjusted according to the dynamic complexity of the data, and the privacy leakage risk caused by the fact that the static encryption intensity cannot adapt to the data dynamics in the prior art is solved.

The invention provides a residual terrain correction estimation method based on PINNs and a residual terrain model, and relates to the technical field of geodetic survey, a data set is randomly divided into a training data set and a test data set, a residual terrain gravity anomaly of each data set is obtained through RTM as an output, the network architecture of the PINNs is a full-connection neural network, and the network architecture of the PINNs is a full-connection neural network. Associating the output quantity and the gravitational potential quantity of the RTM model according to a basic equation of physics geodetics to serve as a boundary value condition; and taking the weighted sum of the physical loss function and the data loss function as a total loss function and an initial full-connection neural network weight parameter of the full-connection neural network, training the constructed full-connection neural network through a training data set, and achieving the requirements of small calculation amount, low required data complexity and high precision through the method.

The invention discloses a large-model multi-agent framework-driven medical data analysis method, which comprises the following steps of: S1, based on collected electronic medical record data of a patient, constructing a medical knowledge graph in an auxiliary manner by virtue of a large language model, and storing the medical knowledge graph in a graph database; s2, constructing a graph retrieval enhancement generation agent, combining the graph retrieval enhancement generation agent with a medical knowledge graph, and performing preliminary analysis on the input data; s3, the assessment agent assesses the data complexity according to the original medical data and the analysis result in S2, and the recruitment agent creates a corresponding number of medical expert agents according to the assessment result; and S4, designing a multi-agent cooperative analysis algorithm, calculating a weighted similarity analysis matrix, judging the consistency of analysis results of the medical expert agent group, and finally obtaining a reliable analysis conclusion. According to the method, the illusion problem of a large model is effectively relieved, and the quality, accuracy and credibility of medical data analysis are remarkably improved.

The invention provides a method and device for analyzing a video and electronic equipment. The method comprises the following steps: acquiring video image data and a video analysis task; performing image sparse processing on the video image data by using a target detection model according to the video analysis task to generate sparse image data; performing feature extraction on the sparse image data by using a visual embedding model to generate visual embedding features; according to a multi-modal cue word template corresponding to the video analysis task, constructing multi-modal dialogue data based on the visual embedding features; and processing the multi-modal dialogue data by utilizing a visual language model to generate an analysis result corresponding to the video analysis task. The analysis result of the visual language model on the multi-modal dialogue data is the result of the corresponding video analysis task, and the data volume and the data complexity needing to be processed in the multi-modal feature extraction process are reduced through sparsification, so that the processing efficiency is improved, and the processing efficiency of the video analysis task is improved by utilizing the visual language model.

The invention discloses a dynamic resource optimal configuration method for multi-source data fusion, and relates to the technical field of optimal configuration, and the method comprises the steps: obtaining multi-source data, including business demand data, resource state data and environmental condition data, and carrying out different processing according to the type of the business demand data; for the first type of service demand data matched with the service demand template, calculating a deviation value of the first type of service demand data relative to a preset service standard, and adjusting a resource allocation direction according to the deviation value; and for the mismatched second type of service demand data, identifying a service operation scene, and adjusting a resource allocation direction or pausing unnecessary resource supply according to the scene and the data complexity, in addition, receiving an instruction for starting a dynamic resource optimization configuration function before data is acquired, the invention also provides an electronic device and a readable storage medium for realizing the method, and the method can effectively fuse multi-source data, realize dynamic optimal configuration of resources, and improve resource utilization efficiency and business operation flexibility.

The invention belongs to the technical field of network security, and discloses a security event association aggregation and maliciousness analysis method. Preprocessing the security event log set to obtain a security event context sequence and an embedded vector thereof; obtaining an initial vector by the embedding vector of the security event context sequence through an encoder; the decoder calculates an attention vector and predicts event probability distribution according to the embedded vector and the initial vector decoded in the previous step; iteratively training until a termination condition is met, and calculating the total attention distribution of each event according to the obtained attention vector and the security event context sequence; an improved DBSCAN method is adopted, and a data set Z formed by total attention distribution of all events is clustered; and extracting a sample from each cluster, and performing event maliciousness analysis according to an LDA topic model. According to the method, the workload of further analysis can be effectively reduced, event sequences with similar context features are clustered, and the data complexity and calculation cost of subsequent analysis are reduced.

The invention discloses a PFMCGAN-DNN-based unbalanced network flow data anomaly detection method, and the method comprises the steps: obtaining an unbalanced network flow data set with marked data categories, segmenting the unbalanced network flow data set to form a training set, a verification set and a test set, and carrying out the preprocessing of the training set, the verification set and the test set; building an improved generative adversarial network model, namely a PFMCGAN model, training the model by using training set data subjected to feature selection, calculating a loss function and updating model parameters through reverse gradient propagation, after training is completed, generating new data by using the PFMCGAN model, and fusing the new data with original training set data to obtain a balanced training set; building a DNN classification model, training the DNN model by using the balanced training set, calculating a loss function and updating parameters; obtaining a classification threshold by using the verification set; and evaluating model performance, inputting test set data into the model to obtain a prediction category of each piece of input data, and calculating a classification evaluation index. According to the method, the feature extraction method and the deep learning model are combined, the data complexity of the model is reduced, and the accuracy of unbalanced data network traffic anomaly detection is improved.

The invention belongs to the technical field of industrial equipment anomaly detection, discloses an unsupervised equipment anomaly detection method oriented to high-dimensional time series data, and aims to solve the problem that a traditional anomaly detection model is difficult to effectively learn and train due to scarcity of abnormal samples and too high multivariate time series data dimensionality in an actual industrial scene. The one-dimensional time sequence data is learned through a plurality of convolutional auto-encoder networks by using an integrated learning strategy, so that the data complexity of model processing is effectively reduced, an attention extrusion and excitation mechanism is fused, the attention of the model on a specific channel is enhanced, the perception ability of the model is improved, and the interference of redundant information is reduced. And finally, integrating the hidden feature information and the reconstruction difference information of the time sequence data of each dimension, and calculating a new reconstruction error as an abnormal score of the sample. According to the method, the high-dimensional time series data-oriented unsupervised equipment anomaly detection precision can be effectively improved.

The invention discloses a multi-dimensional student physical health monitoring method and system, and relates to the technical field of health monitoring, and the method comprises the steps: collecting data to construct a signal subset, carrying out time-frequency decomposition to construct a time-frequency graph matrix, carrying out autoregression modeling to extract an observation matrix and a design matrix, and constructing a frequency feature vector; according to the frequency characteristics of different frequency bands, calculating contribution factors and determining a dynamic health score, calculating a posterior probability of a health state based on a conditional probability of the health score, and calculating a frequency band offset to construct an abnormal frequency band set. According to the method, a health assessment model is constructed by using an observation matrix and a design matrix, so that signal values in each row are associated with lagging historical values, the time dependence of the signals can be captured, the dynamic characteristics of the health signals are reflected, and the mean value, variance, skewness and kurtosis are calculated based on PSD. And high-dimensional spectrum information is concentrated into a feature vector of a fixed dimension, so that the data complexity is reduced, and subsequent classification and health state analysis are facilitated.

The invention discloses a multi-modal data processing method and system based on an attention mechanism, and relates to the technical field of deep learning, and the method comprises the steps: collecting a multi-modal data set, carrying out the multi-scale time sequence calibration through dynamic time warping, and obtaining a time sequence alignment data stream; performing cross-modal semantic association on the time sequence alignment data stream to form a multi-modal feature vector; performing sparse processing on the multi-modal feature vector by using a multi-head self-attention mechanism to generate potential sparse representation; and carrying out coarse graining analysis and fluctuation mode capture on the potential sparse representation, generating a feature sequence length and a variance descriptor, and carrying out spectral entropy calculation to obtain a data complexity score. According to the method, cross-modal semantic association is performed by using canonical correlation analysis, and meanwhile, differential processing is performed on samples with different complexities through the hierarchical adaptive processing model, so that dynamic matching of computing resources is realized, and the resource utilization rate of multi-modal data processing is remarkably improved.

The invention relates to a deep learning cloud particle target detection method based on a dynamic detection head. According to the method, firstly, complexity evaluation is carried out on a cloud particle image data set, the weighted information entropy, the weighted texture complexity and the weighted local standard deviation of each image are calculated, and the comprehensive complexity is calculated based on the complexity indexes. Then, according to the overall complexity grade of the data set and the image data complexity, a detection head selection standard is formulated, and a proper dynamic detection head combination is selected according to the standard; and finally, performing target detection on the cloud particle image data set by using the selected dynamic detection head combination. Through the method, the most suitable detection head combination can be adaptively selected, the target detection precision and efficiency are effectively improved, and the method is particularly suitable for a large-scale cloud particle image data set with complexity difference.

The invention belongs to the technical field of artificial intelligence, and particularly relates to an AI-based file data content accurate and deep analysis and interpretation method, which comprises the following steps: acquiring multi-format file data and file meta-information, constructing an AI analysis network, extracting text semantic vectors, image visual features, table structure information and document layout features, and constructing a multi-dimensional semantic map. According to a user query intention, semantic extension is performed in combination with a domain knowledge base, an enhanced semantic description vector is generated through a graph attention mechanism, semantic reasoning and relation mining are performed by adopting an improved knowledge distillation Transform model, and a deep analysis conclusion is generated through a multi-hop reasoning model in combination with file data complexity, information density and user requirements. And generating a personalized interpretation report in combination with a user role and a task scene, and outputting an analysis result through a visual interface. Therefore, the problems of poor understanding ability, poor file adaptability and the like in the prior art are solved.

The invention provides a fundamental frequency calculation method for adaptive parameter adjustment, which is suitable for power signal frequency estimation under a noise interference or multi-harmonic background. Firstly, voltage and current signals are subjected to analog-to-digital conversion to obtain digital signals, a CIC filter in a non-recursive form is adopted for multi-stage extraction processing, and the data complexity is reduced; sequentially inputting the extracted signals into a plurality of notch filters with feedback structures for harmonic filtering; and iteratively updating the center frequency parameter of the notch filter through a gradient descent method, and when the error function is minimized, the current frequency parameter is the fundamental frequency of the signal. The method has the characteristics of fast convergence, strong robustness and high calculation precision, and is suitable for scenes of electric energy metering, frequency protection and the like.

The invention provides a water supply pipeline leakage detection method and device, computer equipment and a medium, and belongs to the field of pipeline leakage detection.The method comprises the steps that sample operation information of a pipeline and environment information of the environment where the pipeline is located are obtained; mapping the sample operation information and the environment information into a unified vector space by using a knowledge embedding technology, and generating a sample one-dimensional feature vector; constructing a random forest model RF, and initializing the position and speed of a dragonfly population of a dragonfly algorithm DA; optimizing RF parameters through DA iteration; and training the optimized RF model, obtaining a current one-dimensional feature vector of the pipeline, and inputting the current one-dimensional feature vector into the trained RF to obtain a classification result of the leakage degree of the water supply pipeline. Therefore, the data volume is simplified, redundant feature interference is eliminated, the data complexity is effectively reduced, the data processing efficiency and accuracy are improved, and the RF is optimized through the DA, so that the algorithm is easier to deploy.

The invention relates to the field of power grid load prediction, in particular to a smart power grid medium-short-term load prediction method, which comprises the steps of data acquisition and primary processing, historical load data acquisition from a power grid monitoring system, processing of checking missing values, abnormal values and repeated values for the acquired data, and time sequence feature extraction. Calculating an autocorrelation function of the load data, carrying out discretization processing on the meteorological data, then carrying out principal component analysis, and respectively constructing a medium-term load prediction model and a short-term load prediction model; according to the medium-short-term load prediction method for the smart power grid, the data quality is ensured through multi-source data acquisition and comprehensive preliminary processing, and the inherent rule of data is effectively mined and the data complexity is reduced through a multi-feature extraction and dimension reduction method in feature engineering.

The invention provides a map rendering method and device, equipment and a storage medium, and the method comprises the steps: obtaining multiple real-time data of equipment where a client is located in the operation process of the client, and the real-time data comprises map request data; for each performance dimension, at least one sub-score of the performance dimension is calculated according to the real-time data, and the performance dimension comprises the network condition, the equipment performance and the data complexity; obtaining a plurality of existing rendering strategies, wherein each rendering strategy comprises a rendering mode and a performance score calculation mode; for each rendering mode, calculating the adaptability score of the rendering mode according to the calculation mode of the performance score of each performance dimension and the sub-score of the performance dimension, and taking the rendering mode with the highest adaptability score as a target rendering mode; according to the method, switching is carried out on the basis of the current rendering mode and the target rendering mode to achieve map rendering, and the map rendering system has the environment self-adaptive capacity by dynamically sensing data of multiple performance dimensions.

The invention provides an exploration data anomaly detection method and device based on artificial intelligence, relates to the technical field of data processing, and thoroughly gets rid of the constraint of a fixed frequency band on an unsteady logging signal by enabling a frequency band range to be dynamically adjusted along with data complexity through adaptive frequency band division driven by information entropy. Furthermore, high-frequency dynamic features are reserved through normalized logging values, global information is supplemented through window statistics, cooperative information of associated channels is effectively utilized in combination with cross-channel correlation, feature fusion is performed by using frequency band energy vectors, and high-frequency abnormal features are still kept in a remarkable identification degree after fusion. And the suppression by low-frequency backgrounds or statistical characteristics is avoided. In conclusion, the model can accurately distinguish the easily confused exceptions, and the exception recognition precision is improved.