427 results about "Parametric identification" patented technology

The invention relates to the technical field of AI big data, and discloses an intelligent cold chain integrated management method and system based on AI big data, and the method comprises the steps: carrying out the feature data collection of nodes in a cold chain system, obtaining the original data of the nodes, and constructing a cold chain network initial adjacency matrix; inputting a temperature control parameter identification model to carry out graph convolution operation to obtain a temperature control parameter identification result; performing task decomposition and resource allocation calculation to obtain a multi-region collaborative task allocation scheme; performing dynamic load calculation and inter-region task allocation based on a multi-region collaborative task allocation scheme to obtain real-time task execution state data; performing grading processing on the real-time task execution state data and collecting health assessment data of the cold chain system; according to the method, early warning of temperature control abnormity and equipment-level and system-level cooperative control are realized, and a set of complete intelligent cold chain integrated management solution is formed.

Provided is a self-adaptive identification method for nonlinear dynamic parameters of a reducer, which belongs to the design field of a reducer. The method includes: modeling a harmonic reducer corresponding to a flexible joint as a concatemer of a rigid reducer and an elastic torsion spring, and carrying out dynamic theoretical modeling and parameter variable independence processing on the concatemer to form a dynamic equation for parameter identification; giving an optimized motion trajectory to each joint of a robot and controlling the robot to act accordingly, acquiring relevant data needed for parameter identification based on a built-in torque sensor and double encoders inside the joint; using an offline identification algorithm to accurately identify a plurality of dynamic parameters of a collaborative robot considering joint flexibility and friction, and obtaining a minimum parameter set.

The invention discloses a deep learning-based intelligent identification method and system for rock and soil mechanical parameters, and aims to accurately identify the elastic modulus, strength parameter and permeability coefficient of rock and soil and support the decision and design of engineering projects under complex geological conditions. And starting from data demand analysis, determining a parameter identification target and planning a field test layout. A basic data set is constructed by comprehensively collecting and integrating multi-source data including laboratory and field test data, remote sensing information and the like. The data is pre-processed and analyzed to remove noise and explore associations between the data. The method comprises the following core steps: constructing a rock-soil mechanical model by using a deep learning algorithm, and realizing high-precision inversion of parameters by defining and optimizing an objective function. And verifying an inversion result and carrying out uncertainty analysis. According to the method, an analysis result is fed back to engineering design adjustment and decision support, and scheme optimization is achieved. The model and strategy are dynamically adjusted through continuous monitoring and real-time data, and safety and high efficiency of engineering implementation are guaranteed.

The invention relates to a lightweight component surface defect quantitative evaluation method, system and device, and the method comprises the steps: obtaining a multispectral surface image, laser three-dimensional data and ultrasonic detection data of a lightweight component, and carrying out the fusion registration based on a preset multi-sensor synchronization protocol, and obtaining a component joint data set; performing dynamic pruning analysis on the component joint data set to obtain geometric contour features, and performing internal feature extraction on the component joint data set to obtain internal abnormal response features; stress gradient distribution data and volume ratio parameters are carried out on the geometric contour features and the internal anomaly features; and obtaining an original working condition data set of the lightweight component, carrying out service working condition coupling analysis on the multi-dimensional prediction matrix, and carrying out multi-dimensional quality l parameter identification to obtain component comprehensive quality information. According to the method, the morphology mechanical coupling feature extraction precision of composite defects such as cracks and holes can be improved.

The invention discloses a building outer wall leakage detection method and system, and belongs to the technical field of wall leakage detection. The method comprises the following steps: acquiring infrared thermal imaging, laser point cloud and multispectral imaging data, and generating a three-dimensional visual model with temperature information after environment temperature compensation correction, three-dimensional modeling and data fusion; and then extracting temperature characteristic parameters to identify an abnormal region, screening a to-be-verified region in combination with the building construction node coincidence degree and geometric parameters, and finally determining a leakage position through comparison of multispectral characteristic extraction and a preset leakage characteristic library. The system realizes full-process automatic detection by applying the method. According to the invention, temperature signal conduction limitation caused by insulation layer obstruction is broken through, the problem of misjudgment caused by a thermal bridge effect is solved through multi-source data fusion and spectral feature analysis, and the precision and reliability of leakage detection are improved.

Systems and methods of anomaly detection in time-series are disclosed. A time-series dataset is received and a set of segments is iteratively defined from the time-series dataset by identifying a set of changepoints in the time-series dataset based on a changepoint type and a sensitivity parameter, determining whether the set of segments defined by the changepoints satisfy at least one threshold criteria, and modifying the sensitivity parameter when the threshold criteria is not met or outputting the set of segments when the threshold criteria is met. A segment-specific anomaly detection threshold is determined for each segment in the set of segments and a set of anomaly-flagged segments is generated. The set of anomaly flagged segments are generated by an anomaly detection process based on the segment-specific anomaly detection threshold for a corresponding segment. An anomaly-flagged time-series is generated by combining the set of anomaly-flagged segments.

The invention provides an abnormal network traffic identification method based on small sample data enhancement, and the method comprises the steps: S1, converting one-dimensional traffic data into a two-dimensional gray-scale map, and employing a twin network combined with a dynamic attention mechanism to analyze the differences between various types of traffic data pair by pair; parameters of an attention mechanism in the twinborn attention network are analyzed, and inter-class difference features are identified; s2, selecting a training domain by using the inter-class feature difference matrix, introducing a comparative learning method to judge the difference between real data and generated data, and generating an enhanced data set for each class of traffic; s3, after the enhanced data set issued by the central server is received, each edge server uses the enhanced data set and local sample data to train the twinborn attention network in the step S1; due to different local sample data sets, each edge server is trained to obtain a personalized abnormal traffic identification and classification model suitable for local data distribution, and abnormal traffic is effectively classified.

The invention discloses a meta-model-based technical index construction and management method and system. The method comprises the steps of defining index metadata according to standard specifications, project experience and historical data; according to product design requirements, constructing a structured index template library; quoting an index template library, defining indexes and association relationships between the indexes and a demand model, a SysML model, a geometric model and the like, and forming an index library; calling an index suite, and applying the index set to form a plurality of sets of index parameter solutions; calculating and analyzing the index parameters of different solutions so as to realize preliminary balancing of the solutions; according to the changed index parameters, the associated index parameters are identified, and change prompting is carried out; performing tracking, detection and early warning on the index parameters through calculation and analysis of the index parameters; setting an index verification task according to the product research and development process, and managing the execution condition and result of the verification task. According to the technical scheme provided by the invention, full-life-cycle management of the product indexes can be realized.

The invention relates to an intelligent inspection control system and method. The system comprises an environment sensing module, a situation recognition module, a task management and scheduling module, a path planning module, an execution control module and a communication linkage module. The system can collect various environmental parameters of an inspection area in real time, dynamically adjust the priority of an inspection task after identifying an abnormal state, and generate an optimal path to drive a robot to execute response operation; meanwhile, abnormal information, images, videos, sensor snapshots and other contents are uploaded to a background management platform through a communication linkage module, and sound-light alarm equipment, broadcast equipment and other equipment are linked to carry out field response. The control method has task preemption and path reconstruction capabilities, and supports distributed sensing deployment and whole-process execution state monitoring. Compared with the prior art, integrated fusion of environment situation awareness, task dynamic scheduling and response control is achieved, and the adaptive capacity, the response speed and the safety guarantee level of the system to complex scenes are improved.

The invention discloses a driver controller detection system, which comprises an acquisition module, an acquisition module, a processing module, an identification module, a judgment module, an analysis module and a generation module, the acquisition module captures contact information and electromagnetic interference data of contacts in real time by means of a high-dynamic voltage sensor and a broadband EMC probe; the acquisition module is triggered by a locomotive signal and records dynamic working condition data; the processing module aligns the data, performs denoising and extracts fault parameters; the identification module scans the voltage waveform and judges the transient contact failure; the judgment module monitors EMC data and compares an instruction position, and judges misoperation; the analysis module analyzes fault association, judges a composite type and triggers an alarm; the generation module generates a detection report and a log. According to the system, a multi-dimensional monitoring network is constructed under a dynamic working condition, data reliability is guaranteed, faults are accurately identified, a compound fault source is positioned, and the safety and stability of locomotive operation are improved through risk grading early warning and automatic reporting.

The invention discloses a TBM operation parameter intelligent decision-making method based on stratum information dynamic identification, and the method comprises the steps: obtaining rock-machine parameters in a TBM tunneling process, carrying out the preprocessing of parameter information, constructing an original database, building three tunneling working conditions, and carrying out the deep learning of a structure based on a table. A rock mass parameter multi-output prediction model and a tunneling parameter multi-output prediction model are established, tunneling efficiency, construction cost, equipment safety and engineering experience are comprehensively considered, a multi-target optimization algorithm is further constructed, rock mass parameters are predicted through the rock mass parameter multi-output prediction model, tunneling working conditions are identified, and tunneling efficiency is improved. And the tunneling parameter multi-output prediction model is used as a fitting function, the TBM tunneling state is optimized according to actual tunneling working condition requirements, and an operation parameter decision value is given. According to the method, not only are the tunneling rate and the cutter consumption considered, but also the equipment safety, the engineering experience and the like are considered, and the given operation parameter decision value can effectively improve the construction efficiency, reduce the cutter cost and ensure the construction safety.

The invention discloses an under-timing variable working mode parameter identification method and device based on statistical characteristic variable window length, and the method comprises the steps: obtaining and processing a response signal sequence of structural vibration under the excitation of a set environment, and obtaining a processed response signal subsequence in a current sliding window, extracting a time-frequency characteristic matrix from the processed response signal sub-sequence in the current sliding window, extracting a principal component matrix from the time-frequency characteristic matrix by adopting principal component analysis, and analyzing and determining a hysteresis period by adopting an autocorrelation function; according to the time-frequency characteristic matrix, the hysteresis period and the principal component matrix of the current sliding window, estimating to obtain modal parameters of the current sliding window; calculating to obtain the statistical characteristics of the current sliding window and designing a self-adaptive window length adjustment strategy to obtain the window length of the next sliding window; and taking the next sliding window as the current sliding window, and repeating the steps. According to the method, the under-timing variable working mode parameters can be accurately identified.

Systems and methods include extracting item parameters from images of items positioned at a POS system. The item parameters associated with each item are indicative as to an identification of each item thereby enabling the identification of each item based on the item parameters. The item parameters are analyzed to determine whether the item parameters match item parameters stored in a database. The database stores different combinations of item parameters to thereby identify each item based on each different combination of item parameters associated with each item. Each item positioned at the POS system is identified when the item parameters for each item match item parameters as stored in the database and fail to identify each item when the item parameters fail to match item parameters. The item parameters associated with the items that fail to match are streamed to the database thereby enabling the identification of each failed item.

The invention discloses a heavy vehicle axle load identification method based on bridge deflection influence line distribution, and the method comprises the following steps: erecting a monitoring device for vehicle parameter identification on a bridge floor, and arranging a structure displacement detection device at the midspan of a bridge to identify the midspan deflection of the bridge; carrying out on-site vehicle running test to obtain influence lines of small heavy vehicles with different vehicle types and vehicle weights on the midspan deflection of the bridge; performing simulation amplification according to the influence line data of the small heavy vehicle; constructing a heavy vehicle axle load identification model, and performing model training through the obtained influence line data corresponding to the heavy vehicle; and inputting the data of the target heavy vehicle during bridge crossing into the heavy vehicle axle load identification model to obtain axle load data of the target heavy vehicle. The method has the advantages that the vehicle axle load is estimated through on-bridge visual recognition and under-bridge deflection calculation, and the influence of device installation on bridge floor traffic can be avoided; and meanwhile, the recyclable structural displacement detection device can effectively reduce the investment cost of equipment.

The invention provides a method and device for detecting vehicle falling into water, a vehicle and a storage medium, the method is applied to the technical field of vehicle falling into water detection, the method comprises the steps that M falling into water detection parameters are obtained through M collection devices of different types in the vehicle, the falling into water detection parameters are used for representing the falling into water state of the vehicle, and M is a positive integer larger than 1; determining M basic probability distribution values corresponding to the M collection devices according to the M water falling detection parameters; according to the M basic probability distribution values, the falling-into-water confidence coefficient of the vehicle is determined; and when the falling-into-water confidence is greater than or equal to the preset confidence, determining that the vehicle is in a falling-into-water state. According to the method, whether the vehicle falls into water or not can be identified by fusing the falling-into-water detection parameters collected by multiple sensors, the reliability and robustness of vehicle falling-into-water detection can be effectively improved, and misjudgment caused by identification of single equipment is avoided.

The invention belongs to the technical field of parameter identification, and provides a thermal power generating unit electro-hydraulic servo system PID parameter identification method and system, and the method comprises the steps: obtaining the historical operation data of a steam turbine, carrying out the preliminary screening of the data, and constructing a parameter identification data set; according to the mean square error between the actual value and the calculated value of the opening degree of the valve, an objective function of PID parameter identification of the DEH electro-hydraulic servo system is established; setting population initial parameters, and generating a chaotic initial population; based on a parameter identification data set and a target function, introducing an initial population value of the generated chaotic initial population into an improved Bayesian optimization model, and obtaining a preliminary optimization result through Gaussian modeling and kernel function prediction; and taking the preliminary optimization result as input, and utilizing a whale optimization algorithm introducing a Levy flight disturbance mechanism to carry out iterative search for multiple times to obtain a final optimization result. According to the invention, the parameter identification precision and robustness are improved.

The invention relates to the technical field of intelligent calibration, and discloses an intelligent calibration recognition method, device and equipment for an electric screwdriver head of an intelligent electric screwdriver and a storage medium, and the method comprises the steps: carrying out the data collection and feature extraction of the electric screwdriver head, and obtaining a feature data set; writing the equipment identification code and the feature data set into a storage chip of the electric screwdriver; constructing a projection neural network for back propagation training to obtain an intelligent calibration recognition model; when the electric screwdriver head is connected with a host, the equipment identification code and the feature data set in the storage chip are transmitted to the host; inputting the real-time working parameters and the feature data set into an intelligent calibration recognition model for parameter recognition and Kalman filtering calibration to obtain filtered calibration parameters; and judging a deviation value between the filtered calibration parameter and the initial calibration parameter, when the deviation value exceeds a preset threshold value, writing the filtered calibration parameter into the storage chip, and updating the calibration coefficient of the host, thereby solving the problem of inaccurate torsion caused by mixed use of the electric screwdriver heads.

The present invention relates to method and system for travel route optimization. The method includes receiving origin location and destination location via user interface. The method includes identifying candidate travel routes between origin location and destination location based on parameters. The parameters comprise type of vehicle being driven by driver, condition of vehicle, driver preferences, essential stops enroute, historical travel route information, historical accident information, historical crime data, current diversion data, current weather data. The method includes generating score associated with candidate travel routes. The method includes generating summary associated with candidate travel routes using machine learning techniques. The method includes providing generated score and summary of candidate travel routes alongside preview of candidate travel routes. The method includes receiving second input for selection of final travel route from candidate travel routes. The method includes providing turn-by-turn navigation guidelines to navigate vehicle through final travel route via user interface.

The invention relates to a bridge structure modal parameter identification method, which is based on a bridge structure tiny vibration video acquired by a camera. VMD (variational mode decomposition), PBVM (phase-based video motion amplification), Gabor filtering-based phase extraction method GBP (Gabor filtering), an improved clustering algorithm ICA (independent clustering algorithm) and an integrated covariance-driven random subspace recognition (SSI-Cov) and FDD (frequency domain decomposition) algorithm are fused to construct a set of complete non-contact structural modal parameter recognition system. According to the method, effective decoupling of multiple vibration modes of a bridge structure, frequency domain feature extraction of small-amplitude vibration and intelligent and automatic modal parameter extraction are achieved under the conditions that the target frequency bandwidth does not need to be known in advance, environmental noise interference exists and no obvious feature target exists, and the method belongs to the technical field of bridge structure health monitoring.

The invention relates to the technical field of hydrological data processing, in particular to a hydrological data treatment method and system. The method comprises the following steps: monitoring environmental parameters of hydrological equipment in real time, and recording the environmental parameters as equipment environmental data; performing environment abnormal feature detection on the equipment environment data, performing environment influence evaluation on the hydrological equipment according to environment abnormal features, and generating environment influence data of the equipment; performing operation parameter identification on the hydrological equipment according to the environmental influence data of the equipment; and performing deviation comparison on the operation parameters and preset standard operation parameters, and marking operation deviation parameters. Through a data processing technology, a time sequence interpolation technology and a spatial interpolation technology, the hydrological equipment is evaluated by environmental influence, so that the consistency, integrity and availability of hydrological data are enhanced; a distributed storage management mode is adopted, and a hydrological database is constructed, so that the hydrological data management efficiency is improved.

The invention discloses a power transmission tower modal parameter automatic identification method, device and system, and relates to the field of power transmission tower structure monitoring. According to the specific implementation scheme, a plurality of measurement points are arranged on a power transmission tower body in the height direction, and response data a are obtained; based on the response data a, a random subspace method is used for solving and drawing a first stability graph, through first modal data between each first stability point and a nearest point of an adjacent order, fuzzy C-means clustering (Fuzzy C-Means Clustering, abbreviated as FCM) is used for clustering the first modal data, false points are deleted, and a second stability graph is obtained; secondly, extracting stable points, meeting conditions, of each order by adopting two-time hierarchical clustering, and averaging to obtain the frequency, damping and vibration mode of each order; according to the method, the modal parameters of the power transmission tower are indirectly solved by using the data collected by the sensors arranged at the multiple measuring points, and the problems that the modal parameter identification efficiency of the power transmission tower is low and the reliability is poor are solved.

The invention belongs to the field of power distribution network parameter identification, and discloses a power distribution network line parameter identification method and system based on a graph neural network, and the method comprises the steps: employing a unified projection and attention weighting mechanism, obtaining the point features of a power distribution network line, and carrying out the self-adaption integration of multi-source heterogeneous observation data; initial edge features are explicitly constructed in neural network propagation of each layer of graph, node and edge representation is alternately updated, and the line parameter prediction precision is improved; power grid core laws such as Ohm's law and node power injection constraint are converted into a loss function to constrain a training process, consistency constraint of a parameter prediction result and a power system physical law is realized, and interpretability and credibility of the prediction result are improved; and inputting the power grid data into the graph neural network optimization model, and outputting a power distribution network line parameter identification result, thereby solving the problems that in the prior art, dynamic information interaction between nodes and edges is neglected, so that parameter identification is sensitive to network structure change, and physically incredible parameter output is easy to generate.

The invention discloses a ship motion attitude dynamic prediction method fusing CNN and BiLSTM, and the method comprises the steps: S1, obtaining ship navigation data, and carrying out the preprocessing of the ship navigation data; s2, establishing a ship navigation three-degree-of-freedom nonparametric model, and obtaining a regression function of the nonparametric model; s3, establishing a ship motion attitude non-parameter identification prediction model, training the ship motion attitude non-parameter identification prediction model, and optimizing a plurality of hyper-parameters by adopting an improved white shark optimization algorithm to obtain an optimized ship motion attitude non-parameter identification prediction model; and S4, dynamically updating the optimized ship motion attitude non-parameter identification prediction model by adopting an incremental learning method so as to realize dynamic prediction of the ship motion attitude. According to the method, a bidirectional long-short-term memory neural network BiLSTM is adopted as a basic model, a time sequence dependency relationship is more comprehensively captured in combination with past and future information, and an improved white shark optimization algorithm IWSO is adopted to perform hyper-parameter optimization on a non-parameter identification prediction model so as to select an optimal hyper-parameter combination, so that the WCBI-SAT can maintain relatively high prediction performance, and the prediction efficiency is improved. And the method has relatively high generalization ability and stability.

The invention provides a rubber bushing dynamic parameter identification method and device, and the method comprises the following steps: obtaining the test data of a driving end and a driven end of a rubber bushing through a test, and calculating the test dynamic stiffness and the test loss angle according to the test data; establishing a simulation model containing at least one dynamic characteristic parameter to be identified, and generating a result of the same caliber in the model according to a calculation caliber which is the same as that of the test so as to extract simulation dynamic stiffness and a simulation loss angle; aggregating differences between the test dynamic stiffness and the test loss angle and differences between the simulation dynamic stiffness and the simulation loss angle on a preset frequency domain evaluation set, and constructing a target function; and optimizing the dynamic characteristic parameters to minimize the target function to obtain the dynamic characteristic parameters of the rubber bushing. According to the invention, the accuracy and stability of parameter identification in the whole vehicle state are effectively improved.

The invention relates to the technical field of dispensing machine optimization, in particular to an adjusting method and system for self-adaptive optimization of a dispensing machine. The method comprises the following steps: identifying defective dispensing abnormal execution parameters by acquiring a dispensing machine operation log and mapping abnormal execution parameters; then, carrying out curing cross-linking abnormity derivation on defect parameters, quantifying interface bonding force deficiency, and further evaluating the sealing efficiency weakening degree; then, based on the sealing efficiency weakening degree data, interface multi-section curing requirements are analyzed, and then self-adaptive correction of the dispensing process is carried out; and finally, designing automatic firmware through the self-adaptive correction data, and embedding the automatic firmware into a control center to realize self-adaptive optimization of the dispensing machine. According to the invention, the optimization technology of the dispensing machine is optimized, so that the optimization technology of the dispensing machine is more perfect.

The invention discloses a power distribution network short-circuit parameter real-time measurement method, system and device based on PMU cooperation and a medium, and belongs to the technical field of power systems and automation, and the method comprises the steps: collecting PMU real-time measurement data, loading preset parameters, and constructing a PMU cooperation measurement network; carrying out a synchronous sampling algorithm, carrying out comprehensive preprocessing on real-time measurement data, and carrying out state reconstruction and inhibiting the influence of abnormal data according to a state estimator; establishing a short-circuit parameter identification model, and executing short-circuit parameter identification; performing multi-dimensional evaluation on the identification result, and evaluating the real-time performance of the algorithm; and based on the evaluation result, updating the key parameters, and executing adaptive adjustment of the system parameters. According to the invention, through high-precision synchronous acquisition, adaptive signal processing and closed-loop optimization mechanisms, comprehensive improvement of the real-time measurement precision, reliability and adaptability of the short-circuit parameters of the power distribution network is realized.

The invention discloses an aviation food truck docking identification method and system based on feature fusion, and the method comprises the steps: obtaining multi-sensor data of an aviation food truck, extracting cabin door positioning information, and carrying out the feature fusion processing to form a docking-pose feature matrix; performing pattern recognition based on the pose feature matrix to obtain track feature parameters, recognizing a cabin door geometric feature spectrum, generating a cabin door inversion feature library through inversion processing, and performing differential comparison to determine space focus features; identifying a cabin door center position and an obstacle interference position by using a space focus feature, constructing a docking failure preset condition to generate a reverse verification factor, and forming a composite space weight map; acquiring a butt joint gradient information identification position deviation area based on the weight map, and converting the butt joint gradient information identification position deviation area into a path optimization trigger area through deviation conversion parameters; adjusting pose information is formed through the trigger area, a recognition coordinate system is constructed for collision conflict detection, a butt joint recognition result is generated, and safe and accurate butt joint of the aviation food truck and the aircraft cabin door is achieved.

The invention provides a flexible resource parameter identification method based on a KAN neural network, and relates to the technical field of power grid resource parameter identification, and the method comprises the following steps: obtaining historical operation data of power grid resources, and identifying flexible resources in the power grid resources through an analytic hierarchy process; according to the identified flexible resources, constructing a flexible resource digital model through a digital twinning technology, and extracting a sample data set according to the flexible resource digital model; based on the extracted sample data set, constructing and training a resource parameter identification model through a KAN neural network; and acquiring operation data of the flexible resources in real time, and identifying flexible resource parameters by using the trained resource parameter identification model. According to the method, the accuracy of identifying the flexible resource parameters by the model can be improved, more accurate decision support can be provided for power grid operators by identifying the flexible resource parameters, and more reasonable decisions can be made in the aspects of resource scheduling, energy distribution and the like.

The invention relates to the technical field of electric power generation equipment monitoring, in particular to an intelligent multi-point monitoring method and system for a generator set. The method comprises the following steps: acquiring data of a generator set, and dividing the generator set into a primary monitoring sub-region and a secondary monitoring sub-region; deploying multiple types of sensors according to monitoring requirements of key components of each monitored sub-region, and collecting multi-dimensional parameters of the sub-regions in real time; the multi-dimensional parameters of the sub-regions are analyzed in real time according to the multi-dimensional parameters, power generation grid connection abnormal conditions are identified, and multi-level warning and abnormal positioning are carried out; the system comprises a multi-point region division module, a multi-dimensional data monitoring module and a power generation grid-connected abnormity warning module. By means of the mode, the abnormal recognition efficiency and adaptability are improved, and the operation safety and grid connection reliability of the generator set are improved.

The invention discloses a recommendation system-oriented privacy sensitive parameter identification and accurate deletion method, and relates to the technical field of recommendation system model optimization and privacy protection. In order to solve the problems that privacy sensitive parameters in an existing recommendation system model are difficult to accurately position and the model performance loss is too large after deletion, according to the evaluation thought of parameter importance and data feature relevance, importance scores of model parameters on two types of data sets are calculated respectively by dividing a forgetting set and a reservation set, and the privacy sensitive parameters are obtained. The privacy sensitive parameters with the forgetting set importance higher than the reserved set importance are screened out, and mask zero setting processing is carried out. According to the method, the model does not need to be retrained or subjected to weight updating, sensitive parameter recognition and deletion can be completed only through single-time forward propagation, and the original recommendation performance of the model is reserved to the maximum extent while the privacy security of a recommendation system is guaranteed. According to the method, accurate identification of privacy sensitive parameters is realized on a mainstream recommendation model, the reduction range of model recommendation accuracy after deletion is controlled within an acceptable range, and the privacy leakage risk is remarkably reduced.