766 results about "Model prediction" patented technology

This invention discloses an IGBT lifetime prediction method based on a GA-Elman-LSTM combined model. Specifically, the method involves: acquiring IGBT aging data; preprocessing the data; using a GA-optimized Elman neural network to import the training dataset into the GA-Elman model for training; constructing error correction training and validation sets using the trained GA-Elman preliminary prediction model; training the LSTM error correction prediction model; and building a GA-Elman-LSTM error correction combined model based on the GA-Elman preliminary prediction model and the LSTM error correction prediction model for final combined prediction. This invention conveniently and efficiently enables the model to converge to the global optimum, effectively improving the model's prediction accuracy; it also improves the model's prediction stability and reduces prediction risk.

A computer-implemented method of predicting, simulating, or forecasting values of one or more specified subject-related attributes during a clinical trial comprises: receiving input data comprising: a medical history of a subject, the medical history comprising values of a plurality of subject-related attributes of a subject; and data specifying a requested output, the data comprising: the one or more specified subject-related attributes of the subject and a time frame; and applying a trained generative machine-learning model to the received input data, the trained generative machine-learning model configured to generate output data based on the input data, the output data comprising: respective values of the one or more specified subject-related attributes of the subject in the specified time frame.

The application provides a method for predicting the operation time of a ship T-beam welding robot based on machine learning, comprising: obtaining the weld seam meters of multiple types of weld seams planned by T-beam welding robot offline programming software, and recording a data set of actual operation time thereof; dividing the data set into a training set and a test set according to the time dimension; performing normalization processing on the weld seam meters of different weld seam types; using a gradient descent method to solve and construct a linear model of multiple types of weld seams; setting hyperparameters and initial values of model variables; training the model; visualizing the output of the result to evaluate the model; predicting the operation time of a new ship section sample data on the model; constructing a linear model function of multiple types of weld seams through the characteristics of section weld seam sample data, training the weight values of each weld seam type in the model affecting the operation time, and predicting the operation time of a new section in a robot welding station, which can be applied to a robot ship manufacturing workshop and improves the accuracy of operation time prediction of the ship manufacturing workshop.

This application relates to the field of steel production technology, and in particular to a method and apparatus for evaluating the flexibility of a hydrogen-based shaft furnace. The method includes: establishing a time-domain dynamic simulation model of the hydrogen-based shaft furnace to obtain production time-domain data; constructing an initial flexibility model of the hydrogen-based shaft furnace considering the direct reduction iron metallization rate based on the first-order dynamic reaction process of the furnace; adjusting the initial flexibility model using the production time-domain data to obtain a final discretized flexibility model; and integrating the discretized flexibility model into the objective function of the hydrogen-based shaft furnace under real-time electricity prices to obtain the flexibility evaluation result. This solves the problem that related technologies often use steady-state models, which are difficult to characterize the dynamic evolution of the furnace's temperature field, gas concentration field, and reaction process under operating condition disturbances such as load changes, hydrogen supply fluctuations, and changes in raw material properties, easily leading to deviations between model predictions and actual operating conditions.

An operator learning model generator is provided for training a smooth operator learning model for predicting airflow dynamics in a room used by a controller connected to a heating, ventilation and air conditioning (HVAC) system. The operator learning model generator includes an interface circuit configured to receive a training dataset via a network connected to a simulation computer, wherein the training dataset includes solution trajectories of airflow in the room for various times series of control actions given to the HVAC system, a memory configured to store the smooth operator learning model comprising an auto-encoder and a neural ordinary differential equation, the training dataset, and training instructions for the smooth operator learning model, and a processor configured to train the smooth operator learning model stored in the memory, wherein the training instructions comprise a jerk regularization that enforces smoothness of the dynamics predicted by the smooth operator learning model.

This invention relates to the field of data information processing technology, and discloses an intelligent operation and maintenance method and system based on data assets. The method includes: collecting device status data through a distributed IoT sensor cluster; standardizing the raw data and optimizing the data transmission path based on a dynamic topology network; dynamically adjusting the data model structure according to real-time data streams; performing multi-dimensional detection based on user-configured data quality rules to generate a quality detection report; cleaning abnormal data in real time based on the detection results to generate an interactive decision view; predicting operation and maintenance risk events using a multi-algorithm fusion model; and dynamically adjusting the workflow task execution path according to the risk level. This method optimizes the data acquisition and transmission process, improves the real-time performance and accuracy of data quality monitoring, and enhances the intelligence of abnormal data processing and operation and maintenance risk prediction, making operation and maintenance management more flexible and efficient.

This invention discloses a method and system for temperature compensation of fiber optic gyroscopes. It acquires the angular velocity output, temperature, and zero-bias drift value under varying temperature conditions, constructs basic features, temporal memory features, and physical interaction feature vectors, and integrates them into a comprehensive physical information feature vector. This comprehensive vector is used as input, and the zero-bias drift value is used as the training label to train the model until convergence. In the online phase, this comprehensive feature vector is reconstructed in real time and input into the model to predict the zero-bias drift, correcting the real-time output angular velocity. The temporal memory feature introduced in this invention quantifies the thermal accumulation history using statistics from a multi-scale sliding window, eliminating the ambiguity problem of thermal hysteresis mapping and significantly reducing computational load. Simultaneously, the physical interaction feature pre-introduces the nonlinear product term of temperature and rate of change into the model, decoupling linear and nonlinear errors, breaking through the generalization bottleneck of traditional black-box models under small sample conditions, and endowing the model with strong physical interpretability and extrapolation capabilities.

This invention discloses a multi-round adaptive data sharding method for building a learning index. Based on the current linear model prediction, it actively introduces empty slots to adjust the position of newly added data to the model. This accelerates data writing, repairs breaks between data shards, reduces the number of shards, and thus reduces index space overhead. This invention solves the technical problems of high learning index space overhead in existing top-down construction methods due to the large number of models, and low access efficiency in existing bottom-up construction methods due to the compact data shards lacking empty slots and requiring external buffers to support dynamic loads. It also addresses the technical problem of increased shard number and index space overhead in existing bottom-up construction methods when the error threshold is small due to the low linearity of large amounts of data.

This invention relates to the field of artificial intelligence technology in reservoir development, and proposes a method for predicting reservoir hydrocarbon distribution based on joint multidimensional attribute constraints. The method includes: integrating data analysis to determine the main controlling factors; collecting relevant data on the main controlling factors to construct a multidimensional dataset; vectorizing multidimensional parameters and constructing a complete dataset through feature fusion; determining optimization mechanisms and functions to improve model prediction accuracy; and constructing a three-dimensional reservoir hydrocarbon distribution prediction model by training with joint multidimensional physical property constraints and the complete dataset. This invention, by combining multidimensional attribute constraints, improves prediction accuracy. Compared to the 81.25% accuracy of traditional one-dimensional methods, this model achieves an accuracy of 91.36%, meeting engineering application standards and providing strong support for oil and gas exploration and development decisions.

This invention provides a method for simulating and predicting urban bird species diversity based on convolutional neural networks, comprising: acquiring current environmental data of the city to be tested; predicting and generating a spatial distribution map of bird species diversity in the city based on the current environmental data using a target bird species diversity prediction model; wherein, during model training, discrete bird observation vector data is aggregated into a grid using spatial indexing and sliding window techniques, and unique bird species values ​​within the window are deduplicated to generate bird species diversity sample data for model training, and the spatial distribution characteristics of bird species are determined based on historical bird observation datasets; an initial convolutional neural network model is constructed and trained to obtain the target bird species diversity prediction model. This improves the accuracy and adaptability of bird species diversity prediction.

This invention relates to the field of information review and management technology, and discloses a dynamic approval management system for work order video uploads, comprising: a video upload module, a deep model prediction module, a knowledge graph inference engine, a dynamic rule conflict detection module, a manual review workbench, a blockchain evidence storage module, a two-way anchoring module, and a change warning module; the video upload module is used to receive videos associated with work orders; the deep model prediction module is used to identify preset violations frame by frame in the video and output pre-review tags, confidence levels, and heatmaps of suspicious areas. This dynamic approval management system for work order video uploads, through parallel verification of deep model prediction and knowledge graph inference engines and dynamic conflict detection, can automatically identify suspicious videos whose model output is inconsistent with business rules and generate suspicious tags with causal chains, guiding accurate manual review and significantly reducing the workload of ineffective reviews.

This invention discloses a method for classifying consciousness disorders in EEG signals based on dynamic graph convolution and channel attention mechanisms, relating to the field of EEG signal recognition technology. According to the method provided in the embodiments of this invention, a complete closed loop is achieved, covering uploading, preprocessing, artifact removal, segmentation, feature extraction, and model prediction. A dynamic graph convolution modeling method with a trainable adjacency matrix is ​​used to adaptively learn functional connections between EEG channels, overcoming the poor generalization problem of static adjacency matrices. Simultaneously, a joint modeling framework of explicit connectivity (PLV) + implicit connectivity (dynamic graph convolution) is used to more robustly capture cross-channel synchronization patterns under low signal-to-noise ratio conditions.

The present application relates to a dynamic regulation method for probiotic solid-state fermentation of oil tea chaff based on digital twinning, and belongs to the technical field of probiotic solid-state fermentation. The method comprises the following steps: constructing a multi-agent decision model based on a reinforcement learning algorithm, learning the multi-agent decision model, using the multi-agent decision model to optimize control parameters, and controlling the working environment of the probiotic solid-state fermentation tank according to the optimal working parameters. The real-time oil tea chaff fermentation data are compared with the oil tea chaff fermentation data within the preset time, and online optimization is performed according to the comparison result. The present application can deeply integrate multi-scale process mechanism, real-time spatial heterogeneity perception, and dynamic regulation method with self-adaptive collaborative decision-making capability to overcome the defects of the prior art, such as regulation lag, neglect of spatial differences, insufficient model prediction accuracy, and limited optimization decision-making capability, thereby realizing efficient, stable and high-quality production of the probiotic solid-state fermentation of oil tea chaff.

The short-term load forecasting method based on SARIMA-random forest combination model comprises the following steps: grouping the original load data by using a sliding window, decomposing the to-be-tested week-before-next day data set of each group to obtain a trend item, a seasonal item and a residual item; establishing a SARIMA model, predicting the trend item to obtain a preliminary prediction result and a residual; clustering weather factors to obtain similar days, grouping to construct a weather-residual data set and establishing a random forest regression model, learning the influence of the weather factors on the residual, and selecting model parameters by using a grid search method; combining the prediction results of the model, and comparing the influence of weather clustering and residual training on the load prediction accuracy. The method can accurately predict the next day load under the condition that the historical load and weather factors of the to-be-tested day are known, and improves the prediction accuracy.

This application provides a model training method, a speech processing method, an apparatus, an electronic device, a computer-readable storage medium, and a computer program product. The method includes: acquiring a first phoneme sequence sample, a first word sequence sample, and a first alignment relationship between the first phoneme sequence sample and the first word sequence sample; in a first training phase, training an initial prediction model based on the first phoneme sequence sample, the first word sequence sample, the first alignment relationship, the first alignment window, the first phoneme loss weight, and a joint loss function of the initial prediction model; determining a state evaluation index for the initial prediction model; and triggering entry into a second training phase when the state evaluation index meets the phase switching conditions. This application can improve the accuracy of the model's prediction of alignment relationships.

The application provides a new energy electric drive system intelligent torque distribution method based on AI prediction control and related devices; the method acquires vehicle state and sensing data in real time, outputs future speed sequence of the vehicle in the prediction time domain through an environment sensing traffic sequence prediction network; outputs each wheel adhesion coefficient estimation value through a road adhesion coefficient online identification network; constructs a feedforward-feedback dual-channel model prediction controller, solves a feedforward torque distribution sequence based on the future speed sequence in a feedforward layer, and generates a feedback torque correction amount in a feedback layer based on a state deviation; dynamically generates a stability and economy weight vector according to an adhesion margin through a multi-objective dynamic weight distribution network; corrects a total optimization target with the dynamic weight and solves each motor torque instruction; and utilizes an online experience playback buffer to perform incremental online learning on the network during steady-state cruising. The application realizes forward-looking prediction, online identification, dynamic trade-off and continuous evolution, and improves the distribution performance.

The application discloses a fan control parameter online optimization method and system based on digital twinning, which inputs real-time SCADA data into a fast digital twinning model, performs millisecond-level state estimation to meet the real-time response requirement of the control system. At the same time, the slow digital twinning model is driven by buffer data for deep calibration, a high-fidelity state log containing physical mechanisms is generated, and the knowledge distillation technology is used to migrate the physical knowledge of the slow model to the fast model for periodic correction of the fast model. Finally, the modified fast model is used for rolling optimization under the model predictive control framework to output the optimal control sequence. The scheme effectively overcomes the defects of the traditional physical model that the calculation time is too long, and the pure data-driven model that is easy to drift and lacks physical interpretability, while ensuring the millisecond-level response speed, and improving the physical consistency of model prediction and the robustness of long-term operation.

This provides a real-time simulation and model prediction method for arc additive manufacturing based on event sequences. [Solution] In a real-time process of arc additive manufacturing of metal structures, the method includes the steps of: activating units in real time by event sequence and guiding the heat source in real time; setting parameters for the heat source model, thermal simulation, and mechanical simulation, and performing a real-time thermal-mechanical simulation of arc additive manufacturing; and constructing a theoretical model for a simplified calculation of the residual stress field and correcting the calculation model prediction.

This invention relates to the field of electronic component packaging and heat dissipation technology, specifically disclosing a partition-type PCB potting heat dissipation system and method for high-end chip packaging. The system includes a partition-type potting cavity, a multi-zone independent temperature control unit, a colloid flow monitoring unit, and a closed-loop feedback control unit. This invention constructs a closed-loop intelligent control system that deeply integrates real-time monitoring, model prediction, and actuator regulation. The system not only relies on a preset colloid rheology model but also utilizes real-time data such as ultrasonic viscosity monitoring and laser thickness measurement verification, and employs an online parameter identification algorithm to continuously correct the model. This enables the system to adapt to performance fluctuations in different batches of materials and dynamic changes in the process, transforming the potting process from an open-loop, passive physical filling into a closed-loop, predictable, and optimizable intelligent manufacturing process, thereby stably producing high-end chip heat dissipation packages with high consistency and high reliability.

This invention relates to the field of urban flooding prediction technology, and discloses a method, system, and medium for constructing a training dataset for urban flooding prediction. The method includes: acquiring historical rainfall data, performing event-based segmentation, and extracting multi-dimensional features of each rainfall event; calculating the distance matrix between any two rainfall events in each dimension of features; weighting and fusion of the distance matrices to obtain the comprehensive distance between any two rainfall events; then selecting typical rainfall events through cluster analysis to generate a representative rainfall event set; simulating each rainfall event in the representative rainfall event set using a hydrological and hydrodynamic mechanism model to generate flooding depth distribution labels; and pairing the multi-dimensional features of each rainfall event with the flooding depth distribution labels to generate a training dataset. This invention improves the accuracy of urban flooding prediction by constructing a training dataset closely related to physical processes and capable of enhancing the predictive performance of machine learning models.

The application discloses a kind of arid region reservoir water and sand prediction and coordinated regulation potential evaluation method and system, belong to water resources, silt treatment and reservoir optimization scheduling technical field.The technical problems of traditional model prediction accuracy insufficient caused by arid region hydrology data scarcity, water and sand prediction process is fragmented and reservoir desilting and silt reduction and water storage benefit difficult to be synergistically optimized, its gist is that: long short-term memory network is used to predict future runoff sequence by fusing attention mechanism, and the sediment concentration is calculated by combining the water and sand relationship formula;By setting the critical sediment concentration threshold, the high sediment concentration period is identified and the available water for dredging is calculated;At the same time, the flood resource potential is evaluated by combining reservoir capacity and water demand constraints;Finally, the coordinated scheduling scheme is generated by comparing the two potentials.The application realizes the integrated accurate prediction of water and sand process in arid region and the simultaneous quantitative evaluation of regulation potential, and provides intelligent decision support for multi-objective optimization scheduling of reservoir.

The application discloses a microfluidic droplet generation control method, device, equipment and medium, relates to the field of microfluidic droplet generation, and comprises the following steps: first, dividing nodes and judging the flow state of each node according to a limited parameter by using a flow state classifier; predicting the droplet diameter by using a stacking integrated size regression model based on an ideal droplet flow state node; screening candidate nodes based on an accuracy tolerance constraint; determining local sensitivity based on a virtual flow rate micro-perturbation and a flow state boundary risk penalty; determining a robust candidate solution set based on a comprehensive risk score according to the local sensitivity and a confidence interval width; extracting a cluster and a representative solution for each candidate node in the robust candidate solution set; and calculating the two-phase flow control parameters for the representative solution of each cluster in the dimension of the continuous phase flow rate and the dispersed phase flow rate by using a univariate step-by-step double-criterion termination algorithm based on a cascade model, so that the time cost and the material cost are reduced.

This invention relates to the technical field of tire rubber formulation optimization, and in particular to a tire rubber formulation optimization system and method based on deep learning, comprising: a data perception and fusion module, an intelligent prediction and modeling module, a collaborative optimization and decision-making module, and a human-computer interaction and feedback module. By precisely matching with the user's production line, it fundamentally changes the current situation where formulation optimization is divorced from actual production, greatly improving R&D efficiency and the rate of results transformation. Furthermore, by establishing a complete closed loop from virtual optimization to actual production and then feedback learning, the system can continuously absorb new production data and adaptively adapt to changes in the production line status, enabling the model's prediction and optimization capabilities to continuously improve over time.

The invention discloses a field weakening trajectory tracking control method for permanent magnet synchronous motors based on enhanced model predictive current control, comprising: S1, establishing a mathematical model of the permanent magnet synchronous motor in a synchronous rotating coordinate system and determining the optimal operating trajectory in the field weakening region; S2, constructing an extended virtual voltage vector set synthesized from the basic voltage vector, and constructing a virtual voltage vector containing continuous components, excess components, and disturbance components as feedback quantities, which are adjusted by the field weakening current controller. d The shaft reference current limits the virtual voltage amplitude within the inverter output voltage limit; S3, a nonlinear extended state observer based on a nonlinear function is designed to estimate the current state and total disturbance in real time, and correct the current prediction model. The optimal voltage vector is selected through a cost function, and a switching pulse sequence is generated to drive the inverter. This invention can effectively reduce current ripple when motor parameters are mismatched and overcome the deviation of the field weakening trajectory when model parameters are mismatched.

This application relates to a method, apparatus, device, and medium for predicting the lifespan of magnetic liners based on artificial intelligence. It is applied in the field of magnetic liner lifespan prediction technology. The method includes: acquiring initial historical lifespan data and initial customer operating condition data; preprocessing the initial historical lifespan data and the initial customer operating condition data to obtain historical lifespan data and customer operating condition data; analyzing the historical lifespan data and the customer operating condition data according to multiple prediction methods to obtain various lifespan prediction information for the current customer. The prediction methods include empirical formula method, AI model prediction method, layered wear state method, and similar operating condition matching method; verifying the historical errors of various prediction methods based on the historical lifespan data; and fusing the various lifespan prediction information based on the historical errors to obtain target lifespan information. This application has the effect of improving the accuracy of magnetic liner lifespan prediction.

The application discloses a kind of neural network optimization methods for server cluster load prediction, it is related to load prediction and server cluster management technical field, this method includes: real-time acquisition hardware performance counter data of multiple computing units in server cluster and pre-processing;Hardware performance bottleneck is judged based on the data, when existing bottleneck, dynamically trigger the structure of load prediction neural network model Sparse processing;Subgraph splitting is carried out to the model after sparsification, and multiple logical subgraphs with different computing characteristics are obtained;According to the real-time hardware state of cluster and the characteristics of subgraph, the subgraph is scheduled to the corresponding heterogeneous computing unit and executes in parallel;Online calibration mechanism is added to maintain model prediction accuracy, the method is realized by four core modules, using distributed architecture and event-driven communication mechanism, the application realizes the dynamic adaptation of load prediction model and bottom hardware state, improves energy efficiency and resource utilization, and guarantees prediction accuracy.

A steering control device according to an embodiment of the present disclosure comprises a control circuit capable of calculating a target steering angle of a host vehicle on the basis of travel data including point sequence data indicating the center position of the road on which the host vehicle is traveling and data on a plurality of travel parameters including the travel speed of the host vehicle. The control circuit can calculate a plurality of time intervals by dividing the distance between adjacent points among a plurality of points included in the point sequence data by the travel speed, predict future travel data when the plurality of time intervals sequentially elapse by using model prediction control on the basis of the travel data, and calculate a target steering angle on the basis of the point sequence data and the predicted future travel data.

This disclosure provides a training method, apparatus, electronic device, and chip for an order probability prediction model. The method includes: processing an original dataset using a first model to obtain a first dataset, the first dataset including change values ​​corresponding to at least one preset discount data; processing the original dataset using a second model to obtain a base order probability, the base order probability being independent of the preset discount data; constructing a second dataset based on the first dataset and the base order probability; and training a third model using the second dataset to obtain an order probability prediction model, which is used to predict the order probability corresponding to target discount data. This can improve the accuracy of model prediction.