520 results about "Modelling analysis" patented technology

The invention relates to the technical field of model analysis, in particular to a time-space fusion neural network line topology analysis method for a power distribution network. The method comprises the following steps: obtaining original line topology data corresponding to a power distribution network, and carrying out structured disassembly and preprocessing to construct a space-time double graph structure; constructing a bidirectional dynamic feature interaction mechanism based on the space-time double graph structure, performing multi-scale topological feature extraction, and generating a space-time separated feature vector set; performing deep coupling fusion on the feature vector set subjected to time-space separation to generate corresponding unified topological feature representation containing abnormal topology; and constructing a dynamic topology state prediction model based on the unified topology feature representation to optimize a space-time joint loss function and output a corresponding real-time topology connection relationship and an equipment state change trend, and meanwhile, performing dynamic topology reconstruction to generate a current-moment reliable topological graph corresponding to potential branch disconnection and temporary tripping. The topology analysis accuracy of the power distribution network can be improved.

To significantly improve the prediction accuracy, response speed and management efficiency of large river bank slope disasters, an integrated bank slope monitoring method based on multi-parameter collaborative recognition is proposed. The solution includes step S1 of synchronously collecting data on bank slope displacement, pore water pressure, inclination angle, vibration frequency and environmental temperature and humidity to form an original monitoring dataset and construct a multi-parameter collaborative recognition network; step S2 of using a multi-modal data fusion algorithm to generate a fusion data matrix including spatiotemporal correlation features and perform spatiotemporal data alignment and outlier cleansing; step S3 of combining a geomechanical parameter library and a past disaster case library to output a risk level map and perform dynamic risk assessment model analysis; and step S4 of triggering a multi-level early warning mechanism and generating linked control commands including treatment suggestions to perform multi-level early warning and linked control.

The invention discloses a grouting method and system for enhancing strip mine slope stability, and belongs to the technical field of slope protection. The grouting method comprises the steps that early-stage data are pre-collected, and an analysis model is established; grouting is designed according to a model analysis result; a pressure-resistant grouting pipe is installed and connected with a grouting pump; grouting parameters are adjusted in real time according to feedback of the sensor; after grouting, all data in the grouting holes are analyzed; outputting a slurry permeation three-dimensional cloud picture, a stress change curve and a stability evaluation index; and storing all monitoring data and model parameters. A cyclic self-adaptive system is formed, it is ensured that the grouting process is efficient and accurate, the overall stability is improved for heterogeneous rock mass, dependence on a static geological report is reduced, and therefore the problem that grouting parameters depend on engineering experience or conservative of the static report is avoided.

The invention relates to the technical field of building construction monitoring, and discloses a method for evaluating mechanical properties of a steel structure building construction whole process based on digital twinning. The method comprises the steps of establishing a digital twin model fusing multi-source information, and performing real-time linkage with a sensor network arranged on site. Collected data such as deformation, temperature and wind speed are processed through dynamic fusion and an anomaly recognition algorithm, model parameters are continuously corrected, and real-time dynamic high-fidelity mapping of the mechanical state in the construction process is achieved. And based on the updated model, the intelligent analysis module performs cooperative calculation, autonomously identifies construction abnormity, quantitatively predicts potential risks, generates a process optimization decision instruction and feeds back the process optimization decision instruction to a site. According to the method, a closed-loop regulation and control mechanism from data perception, model analysis to decision execution is constructed, and accurate online evaluation of construction mechanical properties and active prediction control of safety risks are realized.

The embodiment of the invention relates to the technical field of data modeling analysis, and discloses a tunnel rock stratum large deformation analysis method and system based on three-dimensional modeling, and the method comprises the steps: collecting initial three-dimensional point cloud data through three-dimensional laser scanning; constructing an initial three-dimensional geologic model containing spatial morphological characteristics and a rock stratum interface geometric topological relation through space coordinate matching and curved surface reconstruction; finite element unstructured mesh generation is carried out on the model, and a rock stratum mechanical analysis model is established in combination with physical and mechanical parameters; inputting construction parameters and process information, activating unit nodes in sequence, and solving a force balance equation to simulate stress distribution; identifying a potential deformation area through iterative calculation, and adjusting boundary conditions to predict a deformation trend; the three-dimensional space distribution cloud atlas, the time evolution curve and the position marks are fused to generate multi-view deformation early warning information, the accuracy and the visualization degree of tunnel rock stratum deformation analysis are improved, and support is provided for construction safety.

The invention discloses a CAD automatic modeling method and system based on parameter driving. The method comprises the steps that model parameter information needing modeling is obtained; generating a CAD (Computer Aided Design) model on the basis of parameter analysis, geometric construction and surface recognition schemes; performing reverse reconstruction of the three-dimensional CAD model based on topology analysis, geometric feature extraction, pattern recognition and parameter relation inference; modeling, optimization and feedback evolution are carried out based on model analysis, feature learning and generation of parametric modeling rules; and according to the obtained data information, CAD automatic modeling based on parameter driving is completed. According to the method, CAD automatic modeling based on parameter driving is achieved, the reliability is higher, the accuracy is better, and the efficiency is higher.

The invention discloses a submarine pipeline state early warning method and system based on scour-vibration coupling sensing, and the method comprises the steps: obtaining the historical marine environment information of a target submarine pipeline region, laying a monitoring sensor array according to the historical marine environment information, collecting multi-source monitoring data, and carrying out the preprocessing of the multi-source monitoring data, thereby forming a multi-source monitoring data set. And constructing a multi-physical field coupling analysis model, inputting the data set into the model to analyze the response condition of the pipeline in the current seabed environment, and obtaining pipeline response simulation information. And analyzing resonance influence caused by environmental scouring in different sections, generating resonance derivative characteristics, and constructing pipeline multi-modal fusion characteristics in combination with monitoring data. Predicting the future response state of the pipeline through the submarine pipeline state analysis model, and carrying out risk early warning prompt; meanwhile, the fatigue life of each section is analyzed based on resonance derivative characteristics, fatigue-prone sections are identified, and maintenance suggestions are generated and pushed. Therefore, the accuracy and timeliness of sensing the state of the submarine pipeline are improved, and the early warning intelligence of the submarine pipeline is effectively improved.

The invention relates to the technical field of text inspection and analysis, in particular to a localized bidding document error inspection method based on a large language model, which comprises the following steps of: constructing a bidding requirement knowledge graph which comprises a plurality of requirement item nodes, the requirement item node attribute comprises requirement content, a chapter to which the requirement item node attribute belongs and an importance level; establishing a preliminary mapping relationship between the requirement item node and the text content set, and generating a structured bidding file parse body; and carrying out double-round model analysis and error checking, generating an error record for the items judged to be not satisfied, and outputting a structured error list comprising the non-satisfied items and reasons of the non-satisfied items. The method improves the accuracy and coverage of error checking, and is especially suitable for checking scenes of bidding and tendering documents with numerous and jumbled contents and various formats.

The invention discloses a security event traceability system and method based on attack chain analysis, relates to the technical field of data security protection, and is used for realizing multi-level perception of attack activities by building a multi-modal killing chain probe matrix as a structured data basis. Modeling analysis is carried out on attack behaviors of different levels through a deep heterogeneous feature extraction architecture, multi-dimensional features are fused into a cross-layer joint feature vector, a 3D attack deduction sand table is constructed according to the cross-layer joint feature vector, the propagation path and probability of APT attack are simulated, a high-dimensional attack curved surface is drawn, and a 3D attack deduction sand table is constructed according to the cross-layer joint feature vector. And key path nodes and potential transverse diffusion directions are identified, and a defense strategy knowledge graph is constructed according to a modeling result. The traceability system breaks through barriers among threat detection, attack understanding and defense response through a stage linkage mode of'probe sensing-intention recognition-path modeling-defense deduction ', and realizes accurate traceability and efficient disposal of security events.

The invention discloses a PC box girder fire damage identification and prediction method and a monitoring and early warning system thereof, and belongs to the technical field of bridge safety risk assessment, and the method comprises the steps: selecting and constructing the characteristics of PC box girder fire prestress loss and bending resistance bearing capacity loss; on the basis of the selected features, a PC box girder prestress loss database and an anti-bending bearing capacity loss database under the fire hazard are constructed; preprocessing various types of original data in the prestress loss database and the bending resistance bearing capacity loss database, converting parameters into formats which can be analyzed by a machine learning model, and dividing a data set into a training set and a test set; based on the training set and the test set, respectively optimizing hyper-parameters of the model corresponding to the prestress loss and the anti-bending bearing capacity loss; and performing performance evaluation, result visualization processing and SHAP interpretability analysis on the optimized model, and applying the model to damage identification and prediction under PC box girder fire. The problem of large prediction error in the prior art is solved.

Hierarchical modelling and advanced feature engineering discover abnormalities in time series data with irregular trends. Data is collected in real time to ensure temporal integrity in the invention. Extraction filters and isolates useful data. Data cleansing removes noise and extraneous data after preliminary analysis identifies patterns and abnormalities. Feature engineering organizes cleansed data for machine learning algorithms. Primary storage stores this data for fast retrieval and extensive trend analysis. Holidays and weekends provide unique patterns in trend analysis. These trends are used to cluster data and create hierarchical predictive models, starting with a first-order model for general trends and increasing in order to refine residuals. Serializing these models improves storage and retrieval. Trend clusters are created from new data points, and algorithms detect pattern deviations. Statistical tests and machine learning classifiers identify anomalies and create alerts and remedial measures. The system monitors and analyzes incoming data to detect anomalies.

The invention provides a dynamic risk prediction and optimization decision-making method and system for a deep foundation pit construction process. A deep foundation pit numerical simulation model is established, and the model is used for simulating soil mechanical behaviors and supporting system response parameters in the construction process; acquiring multi-dimensional real-time monitoring data of the foundation pit, comparing the monitoring data with a numerical simulation result, and calibrating parameters of the deep foundation pit numerical simulation model according to a comparison result; performing time series data modeling analysis on the real-time monitoring data by using a Transform algorithm, capturing possible risk modes in the construction process, and identifying potential risk signals in the data; on the basis of Transform algorithm analysis and numerical simulation feedback results, the foundation pit construction risk state is predicted in real time, and a real-time risk prediction result is obtained; and according to the prediction result, dynamically adjusting the construction scheme to complete decision optimization. According to the method, the risk management and control capability of the deep foundation pit construction process is remarkably improved, and the construction safety and efficiency are greatly improved.

The invention relates to the technical field of genome data analysis, and discloses an intelligent system, method and equipment for assisting multi-step genome data analysis, and the system comprises a dialogue agent which is used for generating a corresponding answer according to a question of a user, or reading an analysis plan file generated by a workflow agent, generating an analysis interpretation text for the analysis plan file; the workflow agent is used for generating a structured task execution plan according to the to-be-executed analysis task and executing the to-be-executed analysis task; and the modeling analysis agent is used for generating a configuration file and a script based on the user request, constructing a model and generating an analysis result corresponding to the user request in combination with the workflow agent. Through multi-agent cooperation, task division and cooperative scheduling are realized, each agent independently completes task planning, execution control, model analysis and other functions, the bottleneck problem of processing of a traditional single model in a complex process is avoided, error accumulation is reduced, and the execution efficiency and stability of the whole process are improved.

The invention relates to a digital archive intelligent processing method, a storage medium and a system, which are suitable for multi-source heterogeneous archive management scenes such as colleges and universities. The method comprises the following steps of: classifying structured and unstructured data such as paper archive scanning pieces and database views, and extracting metadata and entity information by adopting a scanning and OCR (Optical Character Recognition) technology; a complex table and document content are analyzed through a model, semantic analysis (entity recognition, relation extraction and event abstract) is achieved in combination with a language model of a Transform architecture, and a structured report containing a data abstract, an entity relation graph and abnormal annotations is generated. The system is internally provided with a parameter template automatic generation module, supports cross-page content continuous restoration and sensitive data encryption desensitization, and realizes safe sharing through an API interface. The method solves the problems of low efficiency, difficulty in multi-source data fusion and the like of traditional archive processing, improves the automation level and data value mining capability of archive management, and is suitable for intelligent upgrading of complex archive scenes.

The invention provides a whole-house health intervention lighting system and a lighting method thereof, and the system comprises a distributed multi-mode sensing layer which is used for non-inductively collecting the physiological, behavior and environment information of a user through a sensor module integrated in a lighting lamp; the edge calculation and model analysis layer is used for carrying out preprocessing and feature extraction on the data acquired by the distributed multi-mode perception layer and providing a model self-interpretation function; the decentralized federated artificial intelligence decision-making layer is used for constructing a personalized health model based on a federated learning framework, performing health state prediction and generating an intervention instruction; the cooperative multi-target intervention planning layer is used for converting the intervention instruction into a whole-house cooperative illumination control strategy; and the system health and fault-tolerant reconstruction layer is used for monitoring the state of the system and realizing function reconstruction when a fault occurs, and the organic cooperation among the layers realizes the capabilities of non-inductive monitoring, privacy protection, continuous learning and active intervention of the lighting system.

The invention discloses a child brain health intelligent management system and method based on a large model, and relates to the technical field of artificial intelligence and health cross.The system is characterized in that a data set construction module is used for obtaining multi-modal medical data and daily data of each preset child and performing fusion processing to obtain a data set; the model training module is used for training a preset large model based on the data set to obtain a child brain health management large model; and the personalized brain health suggestion generation module is used for generating a corresponding personalized brain health management scheme according to the multi-modal medical data and daily data of the target child by using the child brain health management large model. According to the method, the defects of a traditional child brain health management mode in the aspects of data integration, model analysis, personalized management and the like are effectively overcome, the accuracy, comprehensiveness and personalized level of child brain health management are greatly improved, and more powerful guarantee is provided for child brain health.

The invention relates to the technical field of safety valve fault monitoring, in particular to a safety valve full-life-cycle fault detection method and system, and aims to solve the problems that sudden safety accidents occur frequently due to the fact that a traditional method depends on leakage flow monitoring and cannot capture hidden faults such as valve element jamming and spring fatigue, a power supply system mostly adopts a single power generation mode, and the power consumption is low. In a low discharge scene, the electric quantity of the energy storage module is used up, and monitoring interruption is caused. According to the invention, the problems of safety valve hidden fault monitoring deficiency and unstable power supply are effectively solved. By means of multi-dimensional data collection, spring elasticity attenuation and valve element movement abnormity can be found in advance, and faults are prevented from being evolved into serious leakage. The composite power supply system maintains basic monitoring in the non-discharge stage, energy is quickly supplemented in the discharge stage, and continuous operation of the whole life cycle is guaranteed. A model analysis result can guide maintenance personnel to accurately position a fault part, and unnecessary shutdown maintenance is reduced.

The invention discloses a temperature monitoring control method and system for a communication machine room or a data center machine room, and relates to the field of machine room temperature and humidity monitoring, and the method comprises the steps: S1, selecting a proper temperature and humidity sensor; s2, drawing three-dimensional models of the communication machine room and the data center machine room; s3, determining layout point locations of the temperature and humidity sensors; s4, formulating temperature and humidity control measures; s5, continuous temperature and humidity data acquisition is carried out; s6, collecting and processing the data; s7, setting a temperature threshold value and an early warning mechanism; and S8, carrying out an analog simulation experiment. According to the method, a three-dimensional model analysis process specially aiming at the machine room is arranged, monitoring sensors corresponding to key positions can be specifically configured according to the actual three-dimensional model condition, some existing heat dissipation structures can be optimized according to the three-dimensional model, and in the actual application process, the heat dissipation efficiency can be improved. And a better temperature monitoring effect and a better temperature control effect are achieved.

The invention relates to the technical field of trajectory analysis, in particular to a trajectory aiming trajectory analysis method which comprises the following steps: acquiring wind speed time-varying data and centroid parameters, calculating wind disturbance stress and screening a maximum value, judging a rotating direction trend, executing pitch angle consistency judgment, and calculating an angle adjustment rate to generate boundary parameters. The method comprises the steps of analyzing a target speed trend to judge delay, calculating a deviation trend difference value, extracting wind disturbance sudden change to mark a high-sensitivity zone, judging that deviation coincides with the high-sensitivity zone, and generating a dynamic correction instruction set. Modeling analysis is carried out on a target delay trend by combining a direction included angle and mode difference change in a target speed trend sequence, and a trajectory correction mapping relation is constructed by combining flight parameters and timestamps, so that the adjustability and striking precision of an overall hit path are effectively improved.

The invention discloses a piano playing posture dynamic correction method and system based on multi-dimensional data, and relates to the technical field of posture correction, and the method comprises the steps: collecting the multi-dimensional data of a player in real time through vision, pressure and electromyographic signal sensors, and generating an original data set; performing cleaning, noise reduction and feature extraction on the original data set, generating a preprocessed data set, and extracting biomechanical feature parameters; based on the extracted biomechanical characteristic parameters, a biomechanical model reflecting the posture state and the potential deviation of the player is constructed, and through multi-dimensional data acquisition and fusion, deep learning model analysis and a closed-loop optimization system, the posture state and the potential deviation of the player can be analyzed. The problems that an existing piano playing posture correction method cannot dynamically monitor in real time, accurately recognize posture deviation, provide personalized correction suggestions and continuously optimize the correction effect are solved.

The invention provides an artificial intelligence chip pre-silicon verification system and method, and relates to the technical field of artificial intelligence chips, and the system comprises a model importing layer which is used for importing a current model; the model analysis layer is used for analyzing the current model to obtain a calculation graph; the distributed strategy configuration layer is used for configuring a distributed topology strategy of a hardware simulator of the artificial intelligence chip based on the distributed strategy of the current model, and adjusting a computational graph of the current model to enable the computational graph to be matched with the distributed topology strategy; the performance prediction layer is used for determining the operation performance data of the current model in the artificial intelligence chip based on the operation performance data of each operator; or determining a performance prediction result of the current model in the artificial intelligence chip based on the performance prediction result of each operator; and the operator implementation layer is used for controlling the hardware simulator to run each operator in the computational graph. According to the system and the method provided by the invention, the end-to-end whole-process verification from model input to result output is carried out on the artificial intelligence chip.

The invention discloses a data management system and method based on a large model, and relates to the technical field of big data analysis, and the method comprises the steps: setting a data monitoring management period, enabling the system to collect the stock change information of various types of medicines in a hospital, analyzing the historical warehouse-out data of a department in combination with a time sequence model, and precisely measuring and calculating a first warehouse-out predicted value; and comparing the daily average dose with a standard limited daily dose, deeply mining compatible drug data associated data in an auxiliary drug use scene, analyzing a compatible drug data ex-warehouse trend, and accurately mastering a collaborative consumption rule. By tracking ex-warehouse proportion changes of target drug data and compatible drug data in a historical period, predicting a proportion trend, predicting ex-warehouse quantity in combination with the compatible drug data, dynamically calculating an auxiliary drug use demand, and deeply deconstructing inventory change data, prospective pre-judgment of the inventory change data is realized, hospitals are assisted to optimize an inventory structure, and the workload of the hospitals is reduced. And the accuracy and response efficiency of inventory data management are improved.

The invention discloses an intelligent control method and system for a constant-temperature incubator based on deep learning. The method comprises the following steps: acquiring an equipment state vector in real time; analyzing the equipment state vector by using an improved model, and predicting an environmental parameter demand trend in a culture process to obtain a prediction result; modeling a control strategy into an MDP Markov decision process through a DRL deep reinforcement learning framework according to the prediction result, and dynamically optimizing power distribution of a heater, a refrigerator, a humidifier and a fan by taking minimization of energy consumption and temperature fluctuation as objective functions to obtain an initial control strategy; and adjusting the power of the actuator according to the initial control strategy by using a fuzzy PID algorithm, monitoring sensor data in real time through an isolated forest model, and adjusting the initial control strategy when abnormality is detected to obtain a target control strategy. The environment demand change is pre-judged in advance, and the hysteresis problem of traditional control is solved.

The invention discloses a spiking neural network deployment method and a deployment tool chain based on intermediate representation, and the method comprises the steps: carrying out the model analysis and structure mapping operation of a trained ANN-ONNX model, generating an SNN-ONNX model with a pulse characteristic, carrying out the optimization of a weight parameter through the combination of transfer learning and a precision fine tuning strategy, and carrying out the optimization of the weight parameter. Therefore, the expression ability and adaptability of the converted model are improved, and the reasoning execution performance of the SNN model on a target platform is improved by adopting an optimization means; meanwhile, an SNN-oriented modular operator component library is constructed, and the portability, maintainability and expandability of operators among different hardware platforms are enhanced by adopting an abstract interface and a design mode of specifically realizing decoupling. According to the method, efficient deployment of the SNN model on a target hardware platform is achieved, performance fidelity of the model under the function equivalent condition is ensured, and the method has wide engineering application prospects in the fields of edge calculation, low-power-consumption intelligent terminals, brain inspiration type artificial intelligence and the like.

The invention relates to the technical field of geological informatization, in particular to an artificial intelligence modeling analysis method for a hydrate pilot production data set, which comprises the following steps of: acquiring logging data, lithology data, stratum physical property parameters and natural gas hydrate production dynamic data; screening, cleaning, complementing, de-noising and standardizing are carried out in sequence to obtain an artificial intelligence modeling data set; and establishing a stratum lithology machine learning recognition model, a stratum physical property machine learning recognition model and a natural gas hydrate artificial intelligence historical fitting model through a support vector machine SVM, a random forest RF and a neural network DNN. According to the method, a serial modeling architecture of lithology identification, physical property prediction and production history fitting is created, and the prediction output of the upstream model is used as the optimization input of the downstream model, so that the downstream production prediction model can learn physical property parameters which are recalculated based on machine learning and have higher precision; and the accuracy of final production prediction is improved from the data source.

The invention discloses a distributed power supply-containing power distribution network coordinated scheduling method, which relates to the field of power distribution network coordination, and comprises the steps of S1, multi-flexibility resource refined modeling and interaction characteristic analysis, S2, power distribution network linear load flow calculation and sensitivity analysis, and after the power distribution network linear load flow calculation and sensitivity analysis in S2 are completed, carrying out power distribution network coordinated scheduling. And S3, generating an uncertainty scene based on the generative adversarial network. According to the coordinated dispatching method for the power distribution network containing the distributed power supply, differences of various resources in control modes, response speeds and adjusting capabilities are analyzed through refined modeling, complementary characteristics of the resources are fully utilized by a collaborative optimization model, an overall optimization effect is achieved, a single type of resources are not optimized in an isolated mode, and the method is suitable for large-scale popularization and application. All adjustable resources such as a distributed power supply, an energy storage system, a reactive power compensation device, a flexible load and an intelligent soft switch are incorporated into a unified modeling and optimizing framework, and deep fusion and collaborative interaction of flexible resources can be realized.

The invention relates to a public opinion information intelligent processing method and system based on multi-modal data fusion. The method comprises the steps of obtaining real-time text data, image data, video data and social attribute data; performing keyword rule matching and deep learning model analysis on the text data in parallel, comparing confidence coefficient difference values of rule output and model output, and determining a text analysis result; performing cross-modal fusion on the multi-source feature vectors, and dynamically distributing feature weights according to the data quality of each modal to generate weighted fusion features; calculating a social influence value based on the user authority, the propagation path depth and the time attenuation coefficient in the social attribute data, and identifying a key community cluster; further carrying out weighted matching on a pre-constructed domain strategy matrix through multi-modal evidence; and adjusting a rule weight and a model parameter according to the target domain configuration result and the weighted fusion feature. According to the method, the public opinion field can be accurately identified, the processing scheme is intelligently matched, and a visual and understandable decision basis is provided.

The invention discloses a multi-source heterogeneous big data aggregation analysis method based on master data modeling. The method comprises the following steps: modeling main data, defining a relation mode of a business entity and a data entity, and outputting a graph structure storage mode; classifying the service entities, and defining different data entity fusion algorithms for different data entity types by taking the main attributes of the service entities as the identification basis of the data entities; accessing data, and cleaning the accessed data; service entities and data entities are analyzed from the cleaned data based on a data entity fusion algorithm, and a relation theme library is constructed according to the incidence relation between the service entities and the data entities; storing the cleaned data and a relation theme library thereof; and carrying out modeling analysis on the cleaned data by combining the characteristics of the business entity and the relation theme library. According to the method, unified management of multi-source heterogeneous mass data is carried out from a global perspective, the blood relationship of the data is clear, and the authority is controllable.

The invention discloses a strabismus detection method based on causal feature gating, and the method comprises the steps: carrying out the modeling analysis of a causal relationship between a key point coordinate and a strabismus category through a causal discovery algorithm, obtaining a causal relationship model used for determining the causal relationship, screening the key point coordinate based on the causal relationship model, and obtaining a strabismus detection result. Extracting a plurality of causal variables from a plurality of samples of the training sample set according to the causal variable set; and introducing a causal feature gating module between adjacent network layers of the classification model, training the classification model by using a plurality of causal variables, and performing adaptive adjustment on gating parameters in the training process of the classification model until the classification model converges or reaches a preset training round number, thereby obtaining a final classification model. According to the embodiment of the invention, the accuracy of strabismus classification can be improved, and the interpretability and medical application value of the model can be remarkably enhanced.

The invention discloses a power transformer detection and energy efficiency evaluation method and system based on model analysis, and the method comprises the steps: collecting a performance parameter sequence and an environment parameter sequence in a transformer through a preset frequency, constructing an LSTM prediction model, and carrying out the window training of historical performance data; during prediction, generating a performance change curve and marking an over-threshold boundary segment, synchronously performing piecewise linear fitting on environmental parameters, calculating a fluctuation coefficient and screening an abnormal time period; and defining the intersection of the boundary segment and the abnormal time period as an optimization interval, dynamically adjusting the boundary parameter according to the fluctuation coefficient, generating an optimization parameter sequence, and carrying out precise energy efficiency evaluation through the optimization parameter sequence. Through multi-parameter coupling analysis, the correlation influence of the environmental parameters and the performance parameters is effectively analyzed, the abnormal data state identification capability and the data optimization level of the transformer are improved, the effectiveness and the accuracy of energy efficiency evaluation are improved, and data support is provided for intelligent operation and maintenance of the transformer.