63 results about "Cumulative risk" patented technology

The invention relates to the technical field of computers, and discloses a model content security control method and system, electronic equipment and a storage medium, and the method comprises the steps: combining a received user input with a historical dialogue record of a current session to form a dialogue context; performing risk identification based on the dialogue context, generating at least one risk category and a corresponding initial risk value, and generating a risk integral corresponding to the current input according to a preset integral strategy; accumulating the risk points to accumulated risk points of the current session; comparing the accumulated risk integral with a dynamic risk threshold value obtained by calculation according to a reference threshold value, a logarithmic function attenuation item of the dialogue round and a user historical behavior adjustment item; and when the accumulated risk integral reaches or exceeds a dynamic risk threshold value, triggering a preset risk management and control measure. According to the method, the risk content can be effectively identified and controlled, attacks bypassing a security mechanism through slow induction can be defended, and the security boundary is not easy to detect.

The invention belongs to the technical field of insurance product pricing, and more specifically relates to a multi-factor fusion insurance product dynamic risk pricing optimization method. Collecting multi-source information data, and carrying out data preprocessing; on the basis of travel trajectory analysis of vehicle-mounted OBD data, in combination with high-risk road sections, driving behaviors and environmental factors, a space-time accumulation risk index is constructed, a graph convolutional network is used for modeling a risk transfer effect in a social circle of vehicle owners, mutual influences among the vehicle owners in the social circle are captured, and a neural network based on multi-task learning is used for establishing a social network model. Dynamic risk, social risk features and other related data are comprehensively processed, the risk probability and loss amount prediction are optimized through a shared network layer, and the insurance pricing is generated in combination with external market data, so that the comprehensiveness and accuracy of the insurance pricing are ensured; the method solves a blind area existing in a conventional pricing method, enables the insurance pricing to be more intelligent and precise, and can reflect the actual risk condition of a vehicle owner in real time.

The invention relates to the technical field of medical detection, in particular to a cervical cancer screening system based on risk HPV (human papillomavirus) expansion typing, and a risk assessment module provides solid technical support for individualized CIN2 + / 3 + risk assessment by using a Bayesian network and a Cox proportional risk model. By dynamically acquiring key data such as HPV types, cytological results, patient ages and infection duration, the Bayesian algorithm is beneficial to constructing individual specific risk models, and real-time risk division is carried out. The Cox regression model predicts accumulated risks in three years, dynamically adjusts a risk assessment threshold value and refreshes the weight of an input variable in real time in combination with a machine learning mechanism, and the composite model structure enables the risk assessment to be more flexible and ensures the scientificity of a management decision.

The invention discloses a glioma radiotherapy postoperative risk assessment method based on a magnetic resonance image, particularly relates to the field of glioma radiotherapy patient health risk assessment, and is used for solving the problem that an existing assessment mode depends on manual interpretation and is difficult to predict bad clinical outcomes in advance. The method comprises the following steps: performing clinical data gridding reconstruction on a corticoid use cycle of a patient and tumor molecular typing, and combining morphological characteristics of an edema region in a magnetic resonance image to generate time-aligned clinical comprehensive characteristic vectors; mining a frequent association item set between the comprehensive feature vector and the pathological process to construct a mapping relation model; establishing a probability graph reasoning model of the bad outcome based on the pathological process vector sequence and the probability weight; and integrating the two types of models to form a causal reasoning network, and inputting a target patient feature vector to calculate an accumulated risk value of reaching a bad outcome. According to the method, an interpretable individual risk assessment result can be output, and a basis is provided for postoperative follow-up visit and intervention.

The invention provides a level protection-oriented network security configuration checking method, which comprises the following steps of: constructing a knowledge graph associated with a security control item and a technical implementation mode by relying on a standard document and configuration data, and implementing a control item logic dependency and semantic weight reconstruction model; security configuration of assets is automatically collected and semantically mapped to map nodes, configuration deviation is recognized, a risk influence path template is matched in combination with asset importance and service context, semantic propagation path influence intensity is dynamically calculated, and an accumulated risk influence value of configuration deviation is generated in an aggregation mode; according to the method, risk level judgment is carried out by adopting a piecewise nonlinear function, a structured risk assessment record is output, risk identification and reinforcement suggestion closed loop are realized, and the hierarchical management and control and automatic compliance response capability of security configuration deviation in a network environment is improved.

The invention discloses a blast furnace smelting safety reinforcement learning decision optimization method based on expert strategy constraint. According to the method, an optimization strategy can be learned from an off-line expert track on the premise that operation safety is ensured. Specifically, a conditional generative adversarial mechanism is introduced to realize the alignment of strategy distribution and expert decision, and the exploration ability in a security domain is retained while the expert experience is inherited. And an independent state-action safety evaluation network is designed and a discount factor is introduced, so that the long-term accumulation risk caused by the large hysteresis characteristic of the blast furnace is effectively dealt with. In addition, memory is adopted to enhance network coding historical information, incomplete state observation is supplemented, and information deviation in decision is reduced. According to the method, effective integration of triple guidance mechanisms is realized, knowledge inheritance is realized through distributed alignment, performance improvement is driven through reward optimization, and risk prevention and control are ensured through explicit security constraints.

The invention relates to the technical field of hydraulic engineering safety monitoring and operation control, in particular to a hydraulic engineering information management system based on digital twinning, which comprises a data acquisition module used for acquiring operation state parameters of a hydraulic engineering entity, real-time water level data of a reservoir area and simulation node calculation data generated by a digital twin engine are collected; the first processing module is used for determining a fidelity deviation index between the physical monitoring data and the digital simulation data; the second processing module is used for calculating accumulated risk potential energy reflecting the equivalent exposure duration of the high-risk working condition; the scheduling decision module is used for generating a gate opening degree adjusting instruction for eliminating the accumulated risk through proportional-differential control logic according to the accumulated risk potential energy and a preset safety potential energy threshold value so as to drive a physical gate to act to form a feedback closed loop; according to the method, the technical problem that multi-source heterogeneous data is difficult to fuse and check is effectively solved, data features with small numerical values are prevented from being covered, and the simulation credibility of the system is improved.

The invention relates to the technical field of geographic information systems and space computing, in particular to an amusement equipment safety risk early warning system based on big data. Comprising the following parts: a geospatial data gridding storage module, which is used for collecting multi-dimensional geographic information and space-time state information of an amusement park; the multi-dimensional geographic information refers to space coordinates and proximity relations of amusement equipment; the space-time state information refers to temperature change data in each starting and stopping process of the amusement equipment; and the geographic space correlation degree calculation module is used for calculating an accumulated risk and an instantaneous risk of the amusement equipment based on the multi-dimensional geographic information and the space-time state information. According to the method, geographic space grid storage, dynamic topological association and risk visualization technologies are fused, and intelligent safety monitoring of amusement equipment is achieved; spatial and temporal characteristics are analyzed based on a cumulative / instantaneous risk formula, the early warning capability is improved in combination with closed-loop management of a people flow thermodynamic layer, missed reports are reduced, the operation and maintenance efficiency and the equipment life are optimized, and intelligent decision making is assisted.

The invention provides a mental health risk early warning and intervention system based on multi-modal emotion calculation, and belongs to the field of mental health analysis, and the system comprises a multi-modal fusion module which is used for obtaining multi-modal data, carrying out the constraint and fusion of the multi-modal data, and generating a fused emotion feature; the risk prediction module is used for calculating a mental health risk probability and an emotional state according to the fused emotional characteristics through a two-stage accumulation risk prediction algorithm, and performing mental health risk early warning based on the emotional state; and the strategy recommendation module is used for extracting emotion track characteristics based on the fused emotion characteristics and the mental health risk probability, and carrying out intervention strategy recommendation to obtain an intervention strategy of the current mental health risk. According to the invention, through the multi-modal fusion module, the risk prediction module and the strategy recommendation module, four modal data of video, audio, text and physiology are processed at the same time, a nonlinear accumulation process of mental health risks is accurately modeled, and risk early warning and intervention capabilities are improved.

The present invention relates to the technical field of product lifecycle management, specifically a full lifecycle management system for rope training equipment. The system includes an equipment kit structure definition module, in which an administrator selects independent ropes, quickdraws, belay devices, and safety harnesses, obtains an equipment kit topology diagram, identifies all dependent paths based on the equipment kit topology diagram, calculates a cumulative risk index, and generates a kit critical path list. In the present invention, through the administrator's selection of independent equipment such as ropes, quickdraws, belay devices, and safety harnesses, a topology diagram of the equipment kit with dependency directions is constructed, and all path structures therein are identified. The cumulative risk index is calculated based on the dependency node relationships and risk information in the paths, and a kit critical path list is generated based on this index. This allows subsequent data analysis to focus on the equipment paths that have the greatest impact on system safety.

The present application relates to the field of intelligent medical treatment, in particular to a tumor treatment dosing regimen Pareto optimization method based on forward-backward stochastic differential equation. The present application unifies the terminal tumor load, dosing cost and treatment risk into the same stochastic control framework, can establish the Pareto optimality condition corresponding to the nonlinear model, and further converts into the implementable closed-loop state feedback control law in the LQ case. Through the combination with the numerical example and the Pareto frontier analysis, an implementable, interpretable and scalable technical path is provided for the multi-objective individualized treatment regimen design, which can solve the problem that the existing technology is difficult to uniformly process the terminal therapeutic effect, dosing consumption and cumulative risk, and difficult to obtain the implementable closed-loop feedback strategy under the forward-backward stochastic control framework.

The embodiment of the invention provides a behavior risk prediction method, electronic equipment, a storage medium and a program product. The behavior risk prediction method comprises the steps of obtaining a plurality of static tags and a plurality of pieces of behavior text data of a target object; fusing the behavior text data and a plurality of static tags into a behavior input vector, and encoding the static tags into tag vectors; extracting semantics of each behavior input vector and sorting according to time to obtain a semantic vector sequence; weighting semantic vectors in the semantic vector sequence in each time window in the plurality of time windows, and aggregating weighting results of the time windows to obtain behavior risk representation of the target object; and predicting a risk score of the target object based on the behavior risk representation and the plurality of tag vectors. According to the method, static and dynamic data are fused, short-term burst and long-term accumulation risks are considered through semantic extraction and multi-time window weighting, the dynamism, comprehensiveness and accuracy of risk prediction are improved, and reliable support is provided for risk management and control.

This invention discloses a method, apparatus, medium, and equipment for assessing the transmission of security risks in the software supply chain, belonging to the field of computer software technology. This invention introduces dynamic risk transmission weights and time-series decay coefficients into the assessment of software supply chain security risks. It integrates factors such as component trust level, component dependency depth, and security control measures into dynamic risk transmission weights, accurately simulating the transmission and accumulation process of risks in complex supply chains. Furthermore, the introduction of a time-series decay coefficient precisely calculates the dynamic changes of risks over time. This invention achieves dynamic prediction and time-series modeling of software supply chain security risk transmission, simulating the natural decay of risks over time and providing an assessment synchronized with the actual security situation. The final output cumulative risk value accurately reflects the actual impact on target components, enabling precise location of high-risk indirect dependencies.

The invention relates to the technical field of reliability design analysis of mechanical products, in particular to a non-parametric reliability analysis method for a high-speed driving roller, which comprises the following steps of: S1, acquiring fault data of the high-speed driving roller, and combining the same fault time in the fault data into a fault time point ti; s2, calculating a risk value of each fault time point ti according to a Nelson-Aalen estimation method, and performing continuous accumulation to obtain an accumulation risk function # imgabs0 #; S3, calculating a reliability function # imgabs2 # according to the accumulation risk function # imgabs1 # to obtain reliability; s4, non-parametric reliability degree confidence evaluation based on Nelson-Aalen estimation is carried out; the problem of reliability calculation when the theoretical distribution of the fault time of the high-speed driving roller is uncertain in engineering practice can be solved.

The application relates to an information security detection method and system based on artificial intelligence, a terminal and a medium, and belongs to the technical field of information processing. The information security detection method comprises the following steps: when a transaction event occurs, capturing a transaction key field after data desensitization; generating a behavior vector according to the transaction key field; analyzing a user portrait to obtain user information; loading a corresponding transaction rule set from a local knowledge base according to the user information; obtaining a cumulative risk score of the transaction according to the behavior vector and a set transaction rule in the transaction rule set; judging whether the cumulative risk score is greater than a risk threshold; if yes, triggering an early warning, freezing the transaction and recording an event log; and if no, releasing the transaction. The application has the beneficial effect of comprehensively and objectively evaluating the risk degree of the transaction.

The invention relates to the field of data processing, in particular to a gas pipeline safety installation scheme generation method and system. According to the basic principle of the method, multi-modal data of an installation environment is collected through various sensors; based on the data, utilizing a risk quantification model to respectively generate a risk field aiming at a gas accumulation negative space and a non-uniform repulsion potential field aiming at different types of ignition sources, and fusing the risk field and the non-uniform repulsion potential field to construct a static risk map; on the map, an improved path search algorithm is adopted for pipeline laying path planning, and a weighted cumulative risk cost item is introduced into a cost function of the algorithm to balance safety and economical efficiency; and finally, outputting a safe installation scheme containing the optimized path. According to the method, automation and intellectualization of gas pipeline installation scheme design are achieved, gas accumulation and ignition risks caused by improper pipeline arrangement can be systematically avoided from the source, and the intrinsic safety level and design efficiency are remarkably improved.

The invention is suitable for the technical field of slope monitoring, and provides an intelligent monitoring and early warning method and system for slope stability. According to the invention, stress monitoring of row and column distribution is carried out; force difference analysis is carried out, and the current stage of the target slope area is determined; if the stage is the alarm stage, carrying out slope danger alarm; if the early warning stage is a to-be-early-warned stage, force difference value and force difference area analysis is carried out, and an unstable accumulated value is calculated; and comparing and analyzing the unstable accumulated value, and performing accumulated risk early warning when there is an accumulated risk. The distributed monitoring data can be subjected to force difference analysis, the current stage is determined, when the current stage is the stage to be pre-warned, force difference numerical value and force difference area analysis is performed, an unstable accumulated value is calculated, whether an accumulated risk exists or not is judged, and then accumulated risk pre-warning is performed when the accumulated risk exists. Therefore, the potential slope instability risk can be accurately identified, judged and pre-warned, and the slope safety monitoring requirements of perception in advance and active pre-warning are met.

The application discloses a kind of lack of information mountainous area watershed soil erosion spatial distribution characteristics fast identification method, obtains the multi-source data of study area;By DEM extraction slope factor, by remote sensing image calculation normalized vegetation index and convert into vegetation coverage FVC, by land use data to build natural background mask;Using scoring matrix, erosion risk is calculated year by year, and cumulative risk is obtained by pixel accumulation for many years, and stable high-risk area and deep layer erosion extremely high-risk area are divided;Using 10 Be isotope verification risk area actual sand contribution, through experience relationship to fill in the data of missing branch stream;Constructing step-by-step two-terminal hybrid model, calculate the contribution rate of each branch stream sediment;With sub-watershed as unit, establish risk-contribution double-dimensional coupling model, the matching characteristics of potential erosion risk and actual sediment output are quantified, and finally output sub-watershed division result.The application can quickly identify soil erosion spatial distribution characteristics, key erosion source area and surface deep layer erosion type in mountainous area watershed lacking long-term hydrology sediment data.

The invention discloses a nursing risk prediction system based on artificial intelligence. The system comprises a nursing information acquisition module, a neighborhood accumulated risk construction module, a nursing risk response module, a target design module, a dynamic risk increment mapping module, a model training module and a nursing risk prediction module. The invention belongs to the field of risk prediction, and particularly relates to a nursing risk prediction system based on artificial intelligence. According to the scheme, risk timeliness is reflected through exponential decay weighted accumulation, accumulated nursing risks are smoothly quantified and accumulated through mean neighborhood accumulation, nonlinear mapping and a time interval sensitive kernel are introduced, and nonlinear interaction of nursing information is captured; dynamic risk capture is enhanced through a mean value neighborhood accumulated value, and the sensitivity to instantaneous high fluctuation is improved; filtering slight care fluctuations based on a threshold exempt function; event correlation weighting is introduced, an accompanying acute event is strengthened, risk increment is quantified through a risk development coefficient, and then the nursing risk prediction reliability is improved.

This invention relates to the field of data processing technology, and in particular to a method, apparatus, and storage medium for multi-source alarm data fusion processing in industrial parks. The method includes: constructing a risk propagation network map of the target industrial park based on a topographic map of the park, and setting node vulnerability indices for each risk source; extracting primary events within a monitoring period based on normalized alarm data from each risk source, and extracting diffusion characteristics for each primary event within the monitoring period; performing correlation analysis on the primary events within the monitoring period based on the risk propagation network map of the target industrial park, and constructing a risk causal graph based on the analysis results; correcting the risk causal graph based on real-time environmental data; calculating the cumulative risk value based on the corrected risk causal graph, and generating a multi-level early warning scheme. This invention effectively improves the efficiency of alarm data processing in chemical industrial parks.

The invention discloses a power transmission high-place operation personnel state risk management system and method based on the Internet of Things, and relates to the technical field of electric power safety, and the method comprises the steps: collecting the physiological data of heart rate variability and electromyographic signals during the operation, recording the operation task duration time and the operation type code, and storing the operation task duration time and the operation type code; analyzing the preorder operation load intensity through the load quantification model; continuously collecting data in a rest period, generating a recovery period physiological parameter time sequence curve by taking an operation end moment as a time origin, and calculating a recovery efficiency index; the preorder work load intensity and historical work data are fused, and a personal fatigue resistance portrait is constructed; inputting the load intensity of the preorder operation into the personal fatigue resistance portrait, predicting the fatigue accumulation risk level of the subsequent operation, and setting a recovery efficiency index threshold value; and by comparing the recovery efficiency index with a threshold value, generating an admission instruction and a competent operation type suggestion, or generating a rework risk early warning signal and limiting post admission.

The application provides a user behavior intelligent monitoring method and system based on time axis linkage, and relates to the technical field of user behavior monitoring. The method comprises the following steps: obtaining an initial behavior sequence set of a target user in a monitoring time period; adopting a time sequence analysis method to analyze and determine time sequence dependent characteristics of each behavior segment; determining a cumulative risk dynamic change trend of the behavior segment on a time axis based on the time sequence dependent characteristics; and determining a behavior monitoring result of the target user based on the cumulative risk dynamic change trend. The application can effectively capture the continuity and correlation change of user behavior in the time dimension, quickly perform pattern recognition on abnormal behavior, and improve the accuracy and timeliness of user behavior monitoring.

The invention provides a time sequence disaster data power grid chain generation disaster damage dynamic evaluation method and system, and relates to the technical field of data processing, and the method comprises the steps: carrying out the coupling mapping of a corrected disaster field and power grid equipment, and obtaining an equipment-level dynamic fault probability; processing the equipment-level dynamic fault probability and the power grid topology and operation state data to obtain a disaster evolution event chain; analyzing the disaster evolution event chain to obtain a chain fault accumulation risk index; performing comprehensive assessment according to the chain fault accumulation risk index to obtain a chain disaster damage index; generating an updated original disaster data set according to the newly input time sequence disaster data, obtaining an updated chain disaster damage index, and achieving the dynamic correction of an evaluation result. According to the invention, accurate assessment and real-time correction of power grid chain generation disaster damage are realized, and the timeliness and reliability of disaster risk management are improved.

This disclosure relates generally to method and system for predicting pregnancy complications using a ranked incremental cumulative risk (RICR) scoring. During pregnancy, undetected medical or obstetric events may result in complications that pose risks to both the mother and the neonate. The method displays a questionnaire comprising a plurality of predefined questions on a clinical device to be answered by a subject. In response from the subject a plurality of answers and one or more unanswered questions for the questionnaire are received to compute a plurality of risk score. Then, a ranked incremental cumulative risk score is computed for the plurality of risk scores using a ranked incremental cumulative risk (RICR) scoring technique to generate a rank index. Finally, the risk index, an action plan, a risk tier, and an action plan is displayed on the clinical device to notify a clinician with a machine-generated alert.

This invention relates to the field of wellbore stability monitoring technology in oil and gas drilling engineering. It provides a drilling collapse pressure early warning device based on a three-dimensional geomechanical model, including an early warning decision module connected to the collapse risk assessment module. This module is configured to monitor the temporal changes of the collapse risk index. When the collapse risk index exceeds a first preset threshold a certain number of times within a set time window, a first-level alarm is triggered. When the collapse risk index exceeds a second preset threshold, a second-level alarm is triggered, where the second preset threshold is greater than the first preset threshold. This invention employs a tiered early warning mechanism, setting two alarm thresholds based on the temporal changes of the collapse risk index. This balances the cumulative risk effect with instantaneous abnormal response, enhancing the timeliness and reliability of the early warning, and is suitable for wellbore stability monitoring needs under complex drilling conditions.

The application provides a severe risk prediction method based on a PINN-Cox combined model and related equipment, and relates to the field of intelligent medical information technology. The technology obtains multi-source time series clinical data, and constructs individual survival analysis data sets. On this basis, a PINN-Cox combined prediction model is constructed. The non-linear mapping capability of the deep neural network is used to output a time-dependent relative risk function, which can accurately represent the instantaneous hazard rate in a complex severe scenario. At the same time, the method innovatively defines a mechanism-data combined loss function. In the training, a negative log partial likelihood loss term based on a non-proportional Cox model and a physical residual loss term based on a differential equation of disease evolution mechanism are introduced at the same time, thereby improving the robustness and anti-overfitting capability of the model under the condition of incomplete clinical data such as small samples and irregular follow-up. Finally, the dynamic prediction of the survival function, cumulative risk function and expected remaining life median of the to-be-predicted patient within a preset time window in the future is realized.

A gas turbine blade reliability analysis method uses the distribution of uncertainty variables in the gas turbine blade as input to generate a chaotic polynomial to be solved, obtains the coordinates of sample points to be calculated, obtains the corresponding gas turbine blade wall temperature, calculates the gas turbine blade life corresponding to each sample point to be calculated, obtains the expansion coefficients of the chaotic polynomial, and obtains a response surface equation for the gas turbine blade life; generates the coordinates of sampling points required for fitting a Weibull distribution based on the distribution of uncertainty variables, and calculates the gas turbine blade life corresponding to each sampling point; obtains the parameters required for constructing the Weibull distribution, and obtains the probability density function of the Weibull distribution of the gas turbine blade life; and finally calculates the cumulative distribution function, reliability function, risk function, and cumulative risk function of the gas turbine blade life to achieve reliability analysis of the gas turbine blade.

The application discloses a multi-modal perception-based digital monitoring and intervention system for old people with chronic diseases, and relates to the technical field of health monitoring.The application comprises the following steps: removing artifacts from physiological signals and motion signals, and using a long short-term memory network to construct a personalized physiological baseline model to generate a physiological abnormality deviation index; extracting facial key points from the face image and calculating a psychological state abnormality index, extracting an emotional distribution from the voice signal and calculating a psychological state abnormality index; calculating a mind-body coupling risk value and a cumulative risk index based on the physiological abnormality deviation index and the psychological state abnormality index; and when the cumulative risk index exceeds a preset threshold, outputting a hierarchical intervention strategy in a retrieval-enhanced generation mode based on a pre-constructed double knowledge graph.The application realizes the functional extension from passive monitoring to active intervention through the fusion analysis of multi-modal physiological and psychological data and the quantification of mind-body coupling risk.

The invention discloses a property insurance risk accumulation prediction method based on space grid calculation, and the method comprises the following steps: S1, framing a region range, and determining longitude and latitude points of a starting point; s2, determining a single grid distance according to the analysis elements; s3, drawing a space grid based on the starting point, and carrying out grid coding; s4, vectorizing the target; s5, grid insurance amount calculation; s6, setting a single grid threshold interval; s7, constructing a visual thematic map and different business underwriting decisions based on threshold interval grade division; and S8, according to different service underwriting decisions, realizing insurance service system-level calculation and management and control. According to the property insurance risk accumulation prediction method provided by the invention, accurate analysis of risk accumulation of property engineering insurance of a preserved driver in natural disasters is realized, the defect that the accumulated risk is predicted by administrative division traditionally is effectively solved, the accuracy of the accumulated risk is improved, a basis is provided for underwriting risk control, and high-quality development of financial insurance business is promoted.

The present application relates to the technical field of chip yield prediction, in particular to a GPU chip yield prediction method and system based on multi-dimensional data fusion, comprising: first, obtaining multi-dimensional data in the manufacturing process of the GPU chip; then, based on the process flow, performing time sequence calibration and alignment on the multi-dimensional data, and generating high-order interaction features by using implicit feature interaction technology; then, generating a multi-dimensional abnormal correlation vector by yield bottleneck projection; subsequently, generating a local risk value based on the multi-dimensional abnormal correlation vector and the manufacturing process node, and constructing a global yield map; finally, constructing an abnormal propagation and accumulation calculation framework, calculating the cumulative risk value along the manufacturing process node, and obtaining the final GPU chip yield prediction value through global damage projection. The present application realizes GPU chip yield prediction by performing time sequence fusion and feature cross on multi-dimensional manufacturing data, and constructing a global yield map and calculating the propagation and accumulation risk of abnormalities.