435results about How to "Improve forecast accuracy" 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.

The application discloses a lithology prediction method and device in a target area, equipment, a medium and a product. The method comprises the following steps: acquiring seismic data in a target area, and seismic data and well logging data of known wells in a research area; training an initial well velocity curve prediction model according to the seismic data in the research area and the well logging data of the known wells, to obtain a target well velocity curve prediction model; determining a plurality of virtual well simulation velocity curves according to the seismic data in the target area; inputting the virtual well simulation velocity curves into the target well velocity curve prediction model, to obtain actual velocity curves of the virtual wells; and determining a lithology prediction result in the target area according to the actual velocity curves of the virtual wells and the seismic data in the target area based on an inversion algorithm. The technical scheme can improve the lithology prediction accuracy.

This invention discloses a method for multi-source data fusion and power prediction of wind turbine units, belonging to the field of power prediction technology. It addresses the technical problems of poor multi-source data fusion and power prediction analysis in existing solutions. By filtering and retaining key features through mutual information, redundant information can be effectively reduced. Dynamic weighting via an attention mechanism makes the fused features more adaptable to real-time operating conditions. Differential weight adjustments for wind speed, terrain, and equipment status ensure high relevance of the fused features even in complex scenarios. Spatial feature extraction captures local spatial correlations, temporal features capture temporal dependencies, and random forests handle nonlinear mappings, solving the problem of insufficient generalization ability of single models. The attention mechanism automatically assigns weights to different time steps and features, avoiding complex mathematical processes such as matrix operations and noise covariance estimation.

The application provides a similarity feature constraint-based inverse synthesis prediction method, medium and equipment; wherein the method is: converting a to-be-predicted product into a corresponding product SMILES sequence; using a graph neural network to extract graph structure features corresponding to all characters in the product SMILES sequence; fusing the graph structure features and word vector representations to obtain a fusion vector; inputting the fusion vector matrix into a Transformer encoder and decoder to obtain a reactant set SMILES sequence, and converting the reactant set SMILES sequence to obtain corresponding reactants. The method integrates the graph structure features of the product molecule into the prediction features, improves the prediction accuracy effect, learns the commonality between the reactant and product sets through feature constraint, uses the commonality features to better help the Transformer model to perform inverse synthesis prediction, and further improves the prediction accuracy of the model.

This invention discloses an expert relearning reasoning question-answering method combined with a gating network. It uses a pre-trained language model for scoring and preprocessing to obtain the question context and reasoning graph. Finally, through an expert network, it deeply learns the semantic features of two representation forms and dynamically adjusts the contribution of the two representation forms to answer prediction using a gating network. Experimental results demonstrate that the proposed method is effective for question-answer prediction. By comprehensively considering the different contributions of the question text representation and the reasoned knowledge graph representation to answer prediction, the prediction accuracy can be improved. Experiments on public datasets also show good results, indicating that the method is not only applicable to fixed question-answer prediction but also has a certain generalization ability.

This invention relates to the field of atmospheric particulate matter analysis and discloses a deep learning-based method for calibrating organic aerosol concentrations. This invention addresses the technical problems in existing organic aerosol concentration calibration techniques, such as insufficient utilization of multi-source information, inadequate use of temporal evolution information, and fixed feature fusion methods lacking dynamic weight adjustment. By leveraging multi-source observation vectors, it fully utilizes observation information from different physical properties to establish a stable nonlinear mapping relationship under complex atmospheric conditions and mixed aerosol backgrounds. Through multi-source observation features and historical state features, the model can characterize the dynamic evolution of organic aerosol concentrations. Employing a bidirectional long short-term memory network and SE attention mechanism to process multi-source observation features and historical state features, it can automatically learn and dynamically adjust the importance weights of feature channels, effectively suppressing noise and redundant feature interference. Furthermore, it utilizes forward and backward time-series modeling to fully capture the contextual relationships between consecutive time points.

This invention provides a method for modeling a cylinder motion system based on a physical information neural network, comprising: acquiring time-series data of the cylinder motion system, including time-series control quantities and system state quantities; constructing a physical information neural network model with a dynamic feature input layer and a physical parameter input layer, receiving the time-series control quantities and physical parameter vectors respectively; performing a nonlinear transformation on the physical parameter vectors through a physical parameter encoding network to obtain encoded physical feature vectors; fusing the time-series control quantities and physical feature vectors to form a joint input vector; inputting the joint input vector into the backbone neural network, synchronously predicting the system state quantities through multi-layer nonlinear transformations; constructing a composite loss function including a data loss term and a physical loss term; training the model using time-series data, updating the model parameters by minimizing the composite loss function until the model converges, thereby achieving high-precision modeling using only single motion data.

The present application relates to a kind of lake reservoir type chlorophyll concentration prediction method, device and medium, method includes the following steps: obtaining the water quality, water dynamics and weather of the online automatic monitoring data of lake reservoir including forecast point and upstream point;The online automatic monitoring data is preprocessed, and the data after processing is obtained;The data after processing is carried out feature extraction to obtain the water quality feature and water dynamics index feature of forecast point and upstream point, and construct cumulative illumination feature;Upstream water quality feature, water dynamics index feature and illumination feature are input into the fusion prediction model pre-trained, to obtain upstream transport chlorophyll prediction result and time series influence chlorophyll prediction result, using error reciprocal method upstream transport chlorophyll prediction result and time series influence chlorophyll prediction result are fused;The chlorophyll prediction result obtained by fusion is output.Compared with prior art, the present application has the advantages of high accuracy, strong stability and the like.

The application discloses an energy consumption metering method and device based on digital twinning, which comprises the following steps: collecting building energy consumption related data based on a pre-constructed building energy consumption digital twinning model; analyzing the energy consumption of the building according to the building energy consumption related data, and giving topological correlation characteristics, multi-scale time characteristics, equipment health characteristics and environmental lag characteristics; combining a pre-constructed energy consumption coefficient prediction model to analyze the topological correlation characteristics, multi-scale time characteristics, equipment health characteristics and environmental lag characteristics, and giving environmental influence coefficients, topological correlation coefficients and space utilization coefficients; determining a benchmark energy consumption prediction model through the environmental influence coefficients, topological correlation coefficients and space utilization coefficients, combining an energy consumption compensation coefficient, and giving building target energy consumption, so as to realize accurate prediction of benchmark energy consumption and dynamic compensation of actual energy consumption, realize accurate metering of building energy consumption, and improve the analysis accuracy of building energy consumption.

The application provides a self-adaptive flow control method based on printing content feature analysis, relates to the field of data flow control of a thermal printer, and comprises the following steps: segmenting a to-be-printed data stream and extracting energy density values of each data segment; calculating a virtual heat value in an iterative mode, calculating a physical printing time consumption; calculating a transmission time consumption according to the number of bytes of the data segment and the effective transmission rate of a communication interface; obtaining an actual time consumption of the data segment from sending completion to printing completion, obtaining an error ratio based on the ratio of the estimated time consumption to the actual time consumption, updating an error ratio queue by using a sliding window and calculating the standard deviation thereof, determining a self-adaptive grading threshold value according to the standard deviation, updating a reference coefficient according to the grade to which the error ratio belongs, and dynamically adjusting a heat dissipation attenuation coefficient based on actual time consumption feedback and a virtual heat state. The application realizes accurate prediction and control of the risk of buffer overflow.

The application discloses a heat regulation method for a heat supply system heat station based on secondary side flow, comprising: establishing a load prediction model for each heat station; when the heat load of each heat station changes by more than a threshold value, calculating the required secondary side water supply flow according to the heat load prediction value; establishing a plate heat exchanger model for each heat station heat exchanger; based on a double attention mechanism LSTM model, establishing a data-driven model between the adjustment parameters of the secondary side circulating pump and the auxiliary adjustment device of each heat station and the corresponding data including the primary side water supply flow and temperature, the heat exchange amount between the primary side and the secondary side based on the plate heat exchanger model; adjusting the secondary side circulating pump and the auxiliary adjustment device of each heat station according to the data-driven model to meet the required secondary side water supply flow demand value of each heat station and change the heat entering the secondary side from the primary side; correcting the adjustment parameters of the secondary side circulating pump and the auxiliary adjustment device of each heat station to obtain optimal adjustment parameters.

The application discloses a kind of inventory optimization methods of charging facility spare parts replacement cycle and failure rate analysis, and the core process of this method includes building data layer, multidimensional data is cleaned and optimized, to realize standardized storage;Based on the LSTM algorithm, a multidimensional linkage prediction model is built, the loss effect of design defects and environmental factors is quantified to optimize the model, and the accurate spare parts replacement cycle threshold and equipment failure rate are output;According to the flow properties of spare parts, a differentiated inventory management mode is designed, and the optimal management mode of slow-moving parts is selected;A full-cost quantification model is built to calculate the total cost and service level under each mode;Finally, with the goal of minimizing cost, an intelligent inventory control model is built, which dynamically outputs inventory warning thresholds, replenishment quantities and times;Through the algorithmization and parameter quantification of the whole process, the application realizes fine and intelligent management of inventory, significantly reduces operating costs and stockout risk, and improves the stability of charging facilities.

This specification provides an image processing model training method and apparatus. The image processing model training method includes: adding noise to a target image in an image sample pair to obtain a noisy image; inputting the original image in the image sample pair into an initial image processing model for processing to obtain an initial image; generating a restored image based on the noisy image and the initial image, wherein the original image and the restored image have the same attribute features; training the initial image processing model based on the restored image and the noisy image until a target image processing model that meets the training conditions is obtained, wherein the target image processing model is a machine learning model. When training the initial image processing model based on the restored image and the noisy image, the initial image processing model can gradually incorporate image features of different noisy images during the training process, which can effectively improve the stability of model training, improve the prediction accuracy of the model, and meet the actual image processing needs of users.

This invention belongs to the field of port operation and management technology. It provides a port project full-process decision management system based on big data analysis, comprising: verifying and analyzing multiple historical estimated arrival times using a hypothetical moving window; determining the optimal moving window using the moving window averaging method and Euclidean algorithm; obtaining a model correction factor based on historical estimates and actual times; obtaining the estimated arrival time; comparing it with the waiting-to-departure time; if a mismatch is found, removing the original stable connecting vessels; reselecting vessels based on connection matching degree and re-evaluating; until a vessel with a time-dimensional connection match is found within the connecting vessel sequence. By optimizing the moving window and introducing correction factors, the accuracy of arrival time prediction is improved, avoiding delays due to time mismatches and effectively improving port vessel scheduling efficiency. This invention solves the problem of how to accurately predict the arrival time of stable connecting vessels and determine whether it matches their waiting-to-departure time.

The application relates to a river closure date prediction method based on historical similar year performance driving, which comprises the following steps: using a plurality of feature selection algorithms and leave-one-out cross-validation to screen target predictor sets respectively matched with selected machine learning models and statistical models; based on parameter sensitivity analysis and a Bayesian optimization algorithm, sensitive hyperparameters of the machine learning models are optimized to obtain optimized machine learning models; the statistical models and the optimized machine learning models are configured as candidate river closure date prediction models; a K-neighbor algorithm is used to search a set of similar historical years in a historical observation data set according to current observation data, and a target river closure date prediction model is dynamically optimized according to the comprehensive prediction error of the candidate river closure date prediction models on the set of similar historical years, and a prediction result is output, so that the advantages of multiple models are effectively fused, the generalization limitation of a single model in a complex non-stationary environment is avoided, and the accuracy and robustness of the prediction result are significantly improved.

The application discloses a small-sample steel quality multi-index joint prediction method and device and a medium, relates to the field of material performance prediction, and comprises the following steps: determining a standardized data set comprising steel chemical composition, process parameter characteristics and multi-dimensional mechanical performance index data; configuring a target order of the multi-dimensional mechanical performance index; constructing a regression chain model based on the target order; in the regression chain model, the prediction value of a previous mechanical performance index is used as an additional input feature of a corresponding regressor of a subsequent mechanical performance index; adopting a nested cross-validation strategy to perform hyperparameter search and optimization on the regression chain model, and determining a target hyperparameter combination; performing full-amount retraining on the regression chain model based on the target hyperparameter combination, and obtaining a trained steel performance joint prediction model; and inputting the chemical composition and production process parameters of the steel to be predicted into the steel performance joint prediction model, and obtaining a joint prediction result of each steel performance index corresponding to the steel to be predicted.

The present application relates to the field of advanced auxiliary driving technology, especially to a new personalized lane departure warning method, device and storage medium. The method first acquires real-time data of vehicle operation and determines the point of departure trend occurrence; secondly, vehicle characteristic data and driver characteristic data are acquired, and the Non-stationary Crossformer model is used to predict the lateral deviation trajectory; thirdly, the correction behavior discrimination result and the deviation area are acquired; finally, when the correction behavior discrimination result meets the deviation area judgment condition, it is judged whether the deviation area is greater than the pre-set deviation area threshold, if yes, the warning is executed, otherwise the warning is not executed; the deviation area is surrounded by the horizontal line where the point of departure trend occurrence is located and the lateral deviation trajectory. Compared with the prior art, the present application has the advantages of effectively improving the prediction accuracy of vehicle deviation trajectory and the individualization degree of vehicle deviation threshold setting, thereby reducing the false alarm rate of LDW system.

The present application relates to the technical field of precipitation forecast, solves the problem that the prior art is prone to underestimate the peak value of precipitation or produce false alarm under strong convective weather, and particularly relates to a short-impending precipitation forecast method fusing GNSS and water conservancy rain measurement radar, GNSS observation data and water conservancy rain measurement radar base data in a monitoring area are acquired, pretreated, corresponding atmospheric precipitable water distribution field and radar reflectivity factor distribution field are generated, based on a preset time sliding window, time sequence characteristics of the atmospheric precipitable water distribution field and the radar reflectivity factor distribution field are extracted, and the two are superimposed in channel dimension to construct a multi-channel spatio-temporal fusion input tensor. The present application overcomes the defect that the traditional extrapolation method rapidly decays with the prolongation of the forecast time, significantly improves the prediction accuracy of the rainstorm center intensity and the falling area, reduces the missed report and underestimation of the precipitation event, and solves the problem of low prediction accuracy of the traditional linear extrapolation algorithm under strong convective weather.

The application discloses a power maintenance action prediction method based on target consistency screening and a bidirectional state space, first constructs an observation side video feature sequence, then carries out action semantic token coding, probability weighted mapping and time sequence modeling, obtains action semantic context representation, constructs a joint input vector based on the action semantic context representation, carries out time sequence modeling and distribution construction through a gate recurrent unit, obtains a feature side target distribution, and extracts an observation action representation; subsequently, a future action prediction model based on a bidirectional selective state space is used for prediction, and multiple candidate future action sequences are output; finally, an action side posterior distribution is constructed, a target consistency score is calculated, and the candidate future action sequence with the minimum score is selected as an optimal action prediction result. The application explicitly models statistical correlation and potential target distribution of maintenance actions, introduces a bidirectional selective state space diffusion generation architecture, and realizes accurate and logical prediction of power maintenance actions.

This invention discloses an InSAR interferometric network optimization method based on a multi-factor coherence surrogate model. The method includes: constructing a multi-factor feature dataset containing SAR imagery, baseline data, and soil moisture data of the study area; constructing a candidate interferometric pair set, and randomly selecting a subset of sample interferometric pairs from the candidate set; constructing a coherence surrogate prediction model, which uses the multi-factor feature data corresponding to the sample interferometric pair subset and the ground truth values ​​of sample coherence to train the model parameters; inputting the multi-factor feature data corresponding to each interferometric pair in the candidate set into the trained coherence surrogate prediction model and outputting the predicted coherence of each interferometric pair and the optimized interferometric pair network. This invention improves the stability and accuracy of unwrapping and time-series inversion, while also reducing the difficulty of interferometric processing and coherence calculation under large-scale data, and can support the needs of large-scale, long-term, and near-real-time monitoring.

This application relates to the field of intelligent healthcare, specifically to a data processing method, device, and program product for renal artery FFR assessment. The method includes acquiring continuous temporal frame data from renal artery angiography; selecting frames from the temporal frame data of the renal artery angiography to obtain multiple time series, the multiple time series including keyframes and N valid frames, where N is a natural number greater than 1; inputting the multiple time series into a segmentation model to perform renal artery segmentation to obtain renal artery vessel masks for keyframes and valid frames; registering the keyframes and keyframe masks with the N valid frames and valid frame masks sequentially based on L scales to obtain L*N registered valid frame data; and constructing 2.5D data based on the registered valid frame data and keyframe data to obtain 2.5D data for renal artery FFR assessment. This application has significant clinical value.

This application discloses a defrosting method for vehicle heat exchangers and a training method for a frosting detection model, belonging to the field of vehicle technology. The defrosting method for vehicle heat exchangers includes: acquiring frosting feature data related to frosting scenarios on the surface of the vehicle's heat exchanger; processing the frosting feature data using a frosting detection model to obtain the degree of frosting on the heat exchanger; the frosting detection model is a pre-trained machine learning model; determining whether defrosting conditions are met based on the degree of frosting and the vehicle's location information, and defrosting the heat exchanger if the defrosting conditions are met. This application can accurately detect the degree of frosting on the surface of the heat exchanger.

This invention provides a method for predicting the health trend of hydropower units based on the TCN-BiLSTM-Attention model, relating to the field of hydropower generator sets. The method involves: S1: acquiring the original vibration signal of the hydropower generator set and denoising the original vibration signal using a TTAO-VMD-EWT joint denoising model; S2: obtaining multiple operating state parameters of the hydropower generator set and calculating the Pearson correlation coefficient, Spearman correlation coefficient, maximum information coefficient, and distance correlation coefficient between these parameters and the vibration signal; S3: calculating the fusion evaluation index of each operating state parameter and the vibration signal, selecting key features, and constructing a multi-dimensional feature vector together with the reconstructed vibration signal; S4: constructing a TCN-BiLSTM-Attention prediction model to predict and output the predicted trend of the hydropower generator set's vibration signal. This method improves prediction accuracy, increases prediction efficiency, and enhances the accuracy, stability, and timeliness of hydropower generator set vibration trend prediction.

The application discloses a projection figure distortion prediction compensation correction method of a laser scanning projection system, and comprises the following steps: carrying out laser scanning projection system projection; acquiring projection figure distortion data, establishing a neural network training data set and a test data set; establishing an initial topological structure, weights and thresholds of a BP neural network; optimizing the weights and thresholds of the BP neural network based on a particle swarm optimization algorithm; carrying out deviation compensation on the coordinate positions of each projection point according to the training output of the PSO-BP neural network; selecting feature points in each part of a to-be-projected plane; projecting the feature points by using a distortion deviation correction projector based on the PSO-BP neural network; acquiring feature point projection coordinate values; subtracting the ideal projection point coordinates without distortion from the corrected feature point projection coordinates; judging whether the difference value meets an error interval; if yes, it is determined that the projection coordinate value distortion error has been corrected and compensated; if no, returning to step three for feedback type multiple iterations until the difference value meets the error interval.

This invention discloses a method for predicting the remaining life of brake friction pads, belonging to the field of brake life prediction. The method includes the following steps: S1, obtaining an initial effective feature set based on Pearson correlation coefficient and strong wear correlation features; S2, constructing a multi-physics coupled nonlinear wear degradation model for the friction pads and outputting multi-physics degradation parameters; S3, obtaining the optimal low-dimensional feature set through dual-source fusion; S4, optimizing the hyperparameters of the hybrid basis learner group using a dynamically perturbed improved gray wolf algorithm; S5, constructing a weighted fusion remaining life prediction model and completing accuracy evaluation and error correction; S6, online adaptive updating to achieve real-time remaining life prediction and life threshold warning. Using the above-mentioned method for predicting the remaining life of brake friction pads, high-precision prediction of the entire life cycle of brake friction pads through online dynamic adaptive updating is achieved, significantly improving prediction reliability and engineering application value.

A method, system, device, and readable storage medium for predicting leading-edge bulging cracks are disclosed, relating to the field of geological disaster monitoring and early warning. Specifically, the method includes determining the active thrust of the landslide on the leading edge based on the landslide mass weight, sliding surface dip angle, seepage hydrodynamics, and additional thrust, wherein the additional thrust is generated by external dynamic disturbance. When the detected active thrust is not less than the passive earth pressure limit, the target vertical bulging displacement is determined based on horizontal shear displacement, maximum dilatation angle, geometric correction factor for shear zone thickness, hydraulic correction factor, initial effective stress, real-time effective stress, and dilatation angle stress sensitivity factor. The predicted leading-edge bulging crack result is determined based on the target vertical bulging displacement and the predicted vertical bulging displacement. This application can improve the accuracy and applicability of leading-edge bulging crack prediction.

The application discloses a software upgrade time prediction method, device, equipment and medium. The software upgrade time prediction method comprises the following steps: acquiring a plurality of historical relative time data, wherein each historical relative time data represents the relationship between the time required for software upgrade at the vehicle end and the software data size; obtaining target sequence data based on the plurality of historical relative time data and a plurality of data numbers corresponding to the plurality of historical relative time data; performing data processing based on the target sequence data to obtain a time prediction model; predicting the target relative time data required for software upgrade at the vehicle end based on the target data number by using the time prediction model; and sending the target relative time data to the vehicle end, so that the vehicle end determines the time required for software upgrade based on the target relative time data and the software data size. The time prediction model obtained through data processing can improve the accuracy of the predicted software upgrade time.

This invention discloses a method for temperature control of a rotary lime kiln based on neural network prediction, belonging to the field of industrial process control technology. The method acquires historical process parameter data during the operation of the rotary lime kiln and preprocesses the data; constructs and trains a dual-path neural network prediction model, learning the temporal evolution of the system state and the influence of control quantities on the state through independent paths, and explicitly superimposing the two to predict the operating state parameters of the rotary lime kiln in future control cycles; constructs a proxy air volume based on fan operating parameters, and establishes a dual-cross-limit control mechanism between the proxy air volume and the coal feed rate; based on this, a predictive control optimization model incorporating dual-cross-limit control is constructed, and the optimal control sequence of the control quantities is obtained through a model predictive control solver, and the first control quantity in the optimal control sequence is written into the controller for execution.

This invention belongs to the field of power system data analysis and power consumption forecasting technology, specifically involving a method and system for predicting the power consumption curve of proxy power purchasers in extreme scenarios. The method first constructs a power consumption forecasting model. The model uses an encoder to extract features from multi-source meteorological data of extreme scenarios through supervised contrastive learning, obtaining meteorological feature vectors for the extreme scenarios. A decoder is then used to predict the total daily power consumption based on the extracted meteorological feature vectors. Multi-source meteorological data for the forecast day is acquired and input into the trained power consumption forecasting model to predict the total daily power consumption for the forecast day. Based on the total daily power consumption for the forecast day, the time-of-use power consumption curve for the forecast day is obtained. This invention achieves high-precision total daily power consumption forecasting under extreme scenarios by efficiently extracting meteorological data features through an encoder and achieving rapid and accurate adaptation with a small number of samples through a decoder.