Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

610 results about "Sample mass" patented technology

Universal pointer meter reading method and device

The invention relates to the technical field of reading identification of meters, solves the problems of inaccurate meter reading caused by omission and false detection of OCR (Optical Character Recognition) identification and large sampling sample size caused by improvement of meter reading accuracy in the prior art, and discloses a reading method and device of a universal pointer meter. Comprising the following steps: carrying out scale and pointer rotation coordinate frame detection and OCR character recognition on a dial plate image, and carrying out one-to-one matching on scales and scale values by using a bipartite graph algorithm; completing the completion of the scale values based on a preset scale value database; the average value of the coordinates of the intersection points of the pointer and the extension lines of the scales is used as the coordinate of the rotation center of the pointer, the angle information of the scales and the pointer is calculated, the reading of the meter is calculated according to the included angle between the pointer and the adjacent scales on the two sides, and high robustness can be achieved for shooting angles, image blurring, illumination, meter smudginess and aged meters; and the success rate and the accuracy of reading identification can be greatly improved.
Owner:WEICHENG INTELLIGENT POWER TECH (HANGZHOU) CO LTD

Electromechanical equipment fault monitoring method and equipment combined with adjacent sample analysis, and storage medium

The invention relates to an electromechanical equipment fault monitoring method and equipment combined with adjacent sample analysis and a storage medium, and relates to the field of control equipment monitoring, and the method comprises the steps: interacting with electromechanical equipment, and receiving electromechanical equipment operation control parameters and an electromechanical equipment monitoring feature matrix; performing non-fault state multi-level adjacency search to obtain an adjacency sample tree; recursively aggregating the adjacent sample trees to generate a standard feature matrix of the electromechanical equipment; comparing the electromechanical equipment monitoring feature matrix with the electromechanical equipment standard feature matrix, constructing an electromechanical equipment feature deviation vector matrix, inputting an electromechanical equipment fault probability prediction model, generating an electromechanical equipment fault probability which is greater than or equal to a fault probability threshold, and generating an electromechanical equipment maintenance instruction. Adjacent sample analysis is introduced, the sample size of fault monitoring is enlarged, meanwhile, the judgment accuracy is guaranteed, automation and intelligentization of fault monitoring of the electromechanical equipment are achieved, and the technical effect of improving the fault monitoring efficiency and accuracy is achieved.
Owner:SHANDONG HUAFANGYUN ENERGY SAVING INTEGRATION CO LTD

Aerodynamic force / torque coefficient small sample modeling method based on sign regression

PendingCN119066870AImplement nonlinear dimensionality reductionReduce the number of samplesGeometric CADSustainable transportationFlight vehicleData acquisition
The invention discloses an aerodynamic force / torque coefficient small sample modeling method based on sign regression. The method comprises the following steps: S1, collecting aerodynamic data of a first aircraft; s2, based on the flow state variable, constructing a correlation parameter expression of an aerodynamic force / torque coefficient of the aircraft; s3, performing small sample pneumatic data acquisition on the second aircraft, and optimizing a constant term in the correlation parameter expression according to the small sample pneumatic data acquisition; and S4, performing polynomial fitting on the small sample pneumatic data by using the optimized correlation parameter expression to obtain an aerodynamic force / torque coefficient polynomial expression. According to the method, nonlinear dimension reduction of high-dimensional input quantities is achieved through a correlation parameter method, the high-dimensional input quantities are combined into low-dimensional correlation parameters, a correlation parameter expression can be migrated to other shapes, an aerodynamic model of a new shape can be established through at least three sample points, high interpretability and data extrapolation capacity are achieved, and the method is suitable for being applied to the aerodynamic model of the new shape. And the sample size required by aerodynamic modeling can be greatly reduced, and a scheme is provided for cost reduction and benefit increase of the test.
Owner:NORTHWESTERN POLYTECHNICAL UNIV

Small sample capacity training method based on deep learning

The invention relates to the technical field of deep learning and small sample learning, in particular to a small sample capacity training method based on deep learning, which comprises the steps of 1, cross-domain data adaptation and feature alignment, 2, meta-knowledge distillation and prototype enhancement, 3, attention-guided small sample fine adjustment, and 4, model uncertainty quantification and iterative optimization. According to the small sample capacity training method based on deep learning, through cross-domain feature alignment, meta-knowledge distillation, prototype enhancement and dynamic iterative optimization, the problems of model overfitting and weak generalization ability in a small sample scene are solved, high-precision model training when the sample size is less than or equal to 50 is realized, and the training efficiency is improved. The method is suitable for data scarce scenes such as medical images and minority language processing.
Owner:SUZHOU JIELIXUN INTELLIGENT TECHNOLOGY CO LTD

Soil nutrient content prediction method and system

The invention provides a soil nutrient content prediction method and system, and relates to the technical field of soil nutrient circulation.The soil nutrient content prediction method comprises the steps that to-be-detected soil sample data are obtained, the to-be-detected soil sample data are input into a soil nutrient content prediction model, and prediction parameters of target nutrients in a to-be-detected soil sample are obtained; wherein the soil nutrient content prediction model is a model obtained after training of a joint random forest algorithm and a depth distribution function. And based on the prediction parameter of the target nutrient, obtaining a prediction result of the target nutrient in the to-be-detected soil sample data so as to evaluate change characteristics of the target nutrient in the current to-be-detected soil sample data. On the basis, the problems that in the prior art, in the aspect of deep soil nutrient prediction, the data quality is low, and the sample size is limited can be solved, the effectiveness and accuracy of deep soil nutrient prediction are improved, and the blank in the prior art is filled up.
Owner:INST OF MOUNTAIN HAZARDS & ENVIRONMENT CHINESE ACADEMY OF SCI

Seismic activity parameter determination method and system based on incomplete seismic data

The invention discloses a seismic activity parameter determination method and system based on incomplete seismic data, and relates to the technical field of engineering seismology, and the method comprises the steps: collecting seismic data, and constructing a historical directory; simulating a seismic directory according to the seismic activity parameters and the activity model; similarity comparison is carried out on the simulated seismic catalog and the historical catalog; and determining a seismic activity parameter combination according to a similarity result between the directories. According to the method, the seismic activity parameters are determined based on the incomplete seismic data, and the sample size is increased by coupling various seismic data, so that the defect of a historical seismic directory can be overcome, and the length of available data can be prolonged; a Monte Carlo method is adopted to simulate a seismic directory so as to obtain a characteristic index sample matrix, and Euclidean distance is adopted to carry out mapping similarity analysis on a historical directory and a simulated directory so as to give a parameter optimization result; the earthquake information is utilized as much as possible, and excessive constraints are avoided.
Owner:INST OF GEOPHYSICS CHINA EARTHQUAKE ADMINISTRATION

Genetic facies identification method of logging parameter fusion clustering based on geological constraint

The invention relates to a lithogenous phase identification method based on logging parameter fusion clustering of geological constraints. The lithogenous phase identification method comprises the steps of collecting logging data of a whole well section of a target well and performing normalization processing; performing fusion calculation on the processed logging data based on geological cognition constraints to obtain fused logging data; clustering the fused logging data by adopting an unsupervised machine learning algorithm and then combining the fused logging data to obtain an initial logging facies type; screening the initial logging facies types based on geological law constraints, and then carrying out intersection merging processing to obtain a target logging facies type; and a plurality of sample points are selected in the target well coring well section, lithogenous phase marking is carried out, a lithogenous phase data set is constructed, target well logging phase types are calibrated and verified, the lithogenous phase type corresponding to each target well logging phase is obtained, and lithogenous phase recognition of the whole well section is completed. According to the method, various logging parameters are comprehensively applied under geological constraints, and accurate identification and prediction of the lithogenous phase can be realized under the condition of low sample size.
Owner:CHINA UNIV OF PETROLEUM (EAST CHINA)

Large model optimization method and device and medium

The invention provides a large model optimization method and device and a medium, and relates to the technical field of computers, and the method comprises the steps: obtaining a basic large model pedestal which comprises a basic large model and a basic reward model for rewarding the output of the basic large model; forming generalization data by using high-quality samples of a plurality of professional scenes of the target domain, and performing generalization training on the basic large model to obtain a generalization large model of the target domain; using high-quality samples and non-high-quality samples of each professional scene in the target field to form professional data, and performing vertical training on the basic reward model and the generalization large model to obtain a special large model of each professional scene in the target field; and deploying the special large models to each professional scene of the target field for application. The generalization ability of the large model is improved by adopting the thought of meta-learning, the available sample size during vertical training is increased by adding the reward model, and the problem of training imbalance caused by small high-quality sample size is avoided.
Owner:CHINA UNITED NETWORK COMM GRP CO LTD

Farmland soil sampling space optimization method based on local heterogeneity

The invention discloses a farmland soil sampling space optimization method based on local heterogeneity, and belongs to the field of soil sampling space optimizing.The method comprises the steps that the heterogeneity level of soil in a research area is calculated through a variable coefficient; according to the soil heterogeneity level of the research area, dividing the research area into spatial continuous sub-areas with similar soil heterogeneity level by using a heterogeneity change algorithm; and on the basis of the environment covariables, a method of combining a fuzzy c-mean classifier and a heterogeneity change algorithm is adopted, geographic and feature spaces are optimized, key sampling positions in continuous sub-regions of each space are identified, and farmland soil sampling space optimization is completed. According to the method, the problem that the influence of the soil local heterogeneity level on sample size distribution and spatial layout optimization is not considered in an existing method is solved.
Owner:ZHENGZHOU UNIV

Micro-fluidic chip and detection method thereof

The invention relates to the technical field of biomedicine, and provides a micro-fluidic chip and a detection method thereof.The micro-fluidic chip comprises a substrate, the substrate is provided with at least one micro-fluidic channel, one end of the substrate is provided with a whole blood treatment area, and the micro-fluidic channels are communicated with the whole blood treatment area and extend towards the other end of the substrate; the whole blood treatment area is provided with a filtering piece, and the filtering piece is used for filtering a sample in the whole blood treatment area; a sample adding hole is formed in the outer vertical face of the base plate, and the sample adding hole is communicated with the whole blood treatment area and used for adding a sample into the whole blood treatment area; the microfluidic channel is sequentially provided with a sample reaction area, a sample detection area and a sample recovery area along a direction deviating from the whole blood treatment area, the sample reaction area is communicated with the whole blood treatment area, the sample reaction area is communicated with the sample detection area through a first capillary channel, and the sample detection area is communicated with the sample recovery area through a second capillary channel. The method has the advantages of being easy to operate, low in sample size requirement, rapid in multi-joint detection and wide in application scene.
Owner:SHANGHAI SCIZENG MEDICAL TECH CO LTD

Material structured data-oriented generation and screening method and device

The invention discloses a material structured data-oriented generation and screening method and device. The method comprises the steps of S1, acquiring material structured data; s2, according to the material structured data, judging whether treatment is needed or not by evaluating the ratio of the characteristic quantity to the sample quantity and the insight of a data set, and if treatment is needed, performing the step S3; s3, effectively generating the structural data of the material by using the depth generative model oriented to the structural data; s4, the generated samples are evaluated through multiple dimensions and weighted integration is carried out; and S5, according to an evaluation result, performing multi-dimensional high-quality sample screening by fusing domain knowledge. By adopting the technical scheme of the invention, effective generation of high-quality material structured data is realized.
Owner:SHANGHAI UNIV

Small sample HRRP radar target identification method, apparatus and device, and medium

The invention relates to a small sample HRRP radar target identification method, device, equipment and medium, after normalization processing is carried out on a one-dimensional high-resolution range profile sequence of an unknown target with a small sample size, radar scattering physical prior is utilized to construct a static adjacency matrix, then a dynamic adjacency matrix is generated through a feature enhancement network, and the dynamic adjacency matrix is obtained. And after the static adjacent matrix and the dynamic adjacent matrix are fused, a hierarchical adaptive meta-learning graph neural network is adopted to carry out target identification training, so that high-precision identification of a position target is realized under data of fewer samples.
Owner:NAT UNIV OF DEFENSE TECH

Low background liquid scintillation spectrometer

The invention provides a low-background liquid scintillation spectrometer which comprises a packaging shell, and a detection mechanism is installed at the bottom in the packaging shell and used for detecting a to-be-detected sample in a sample bottle; the sampling mechanism is mounted at the top in the packaging shell and is used for transferring a sample bottle to the detection mechanism; the sample bearing frame is located between the detection mechanism and the sampling mechanism, and the sample bearing frame is provided with a sample groove for placing a sample bottle; the additional sample loading mechanism is mounted on the sample bearing frame, is pulled in and out of the packaging shell, and is provided with a sample groove for placing an emergency sample injection bottle. The low-background liquid scintillation spectrometer integrates the functions of automatic sample introduction, automatic sample adding, automatic closing of a lead chamber, automatic sample loading and exiting, automatic static electricity removal of a sample bottle, automatic temperature control of a sample and a detection area, urgent detection of any sample position, automatic sample adding and the like of a large sample size, and is high in automation degree and compact in structure.
Owner:HUBEI FANGYUAN ENVIRONMENTAL PROTECTION SCI & TECH CO LTD

Method for testing tensile strength of injectable hydrogel wound dressing

The invention relates to the technical field of material physical property testing, and discloses an injectable hydrogel wound dressing tensile strength testing method, which comprises: establishing a correlation model of the mass and the maximum fracture load of a standard sample at different temperatures, and the correlation model of the initial rigidity and the temperature; measuring the mass and the actual maximum fracture load of the to-be-measured sample, and determining the actual initial rigidity of the to-be-measured sample; according to the method, the actual initial stiffness is calculated, the equivalent temperature of the test is converted by utilizing the actual initial stiffness, the standard theoretical maximum fracture load is calculated in combination with the sample mass and the corresponding model, and the performance is judged by comparing the two, so that the measurement of the geometric dimension of an irregular sample is avoided by constructing the correlation between the mass and the mechanical performance; the quality control problem of an in-situ forming material is solved, and the reliability of a judgment result in an industrial environment is further ensured by utilizing mechanical response data of the same test through an integrated temperature self-adaptive calibration mechanism.
Owner:陕西扶特林生物科技有限公司

Full-automatic dyeing system and dyeing process

The invention discloses a full-automatic dyeing system and a dyeing technology.The dyeing system comprises a processing cavity, an incubation area is arranged in the processing cavity, the incubation area is provided with a plurality of incubation positions, the number of the incubation positions is n, the dyeing system further comprises a transfer waiting area, the transfer waiting area is arranged in the processing cavity, and the transfer waiting area is arranged in the processing cavity; the transfer waiting area is used for placing slides to be processed and slides to be sent out after processing, before the full-automatic dyeing system starts to work, the number of the slides in the transfer waiting area is N, and N / n is larger than or equal to 1; the reagent bin is arranged in the treatment cavity, a refrigeration mechanism is arranged in the reagent bin, and a plurality of small reagents can be contained in the reagent bin. Through the arrangement of the transfer waiting area and the reagent bin, the off-machine time of workers is prolonged, samples can be continuously processed under the unattended condition at night, the sample processing amount per day is increased, and therefore the high-throughput effect is achieved.
Owner:SUZHOU CHENGYIN MEDICAL TECH CO LTD

Cloud top height quality inspection method and device, computing equipment and storage medium

The invention discloses a cloud top height quality inspection method and device, computing equipment and a storage medium, and belongs to the technical field of atmospheric detection and remote sensing, and the method comprises the steps: taking laser radar cloud profile data as a detection truth value, taking imaging type radiometer cloud product data as to-be-inspected data, and carrying out the quality inspection of the to-be-inspected data on the basis of the time-space matching of the two types of data; the method comprises the following steps: screening out unqualified data through a space-time uniformity test, and carrying out layer-by-layer and step-by-step analysis on qualified matching data to count a total sample size, correct detection, missing detection, false detection, a correct screening sample size and a proportion, a correct identification sample size of each cloud type including cirrus cloud and lower cloud, an error identification sample size and a corresponding cloud top height difference; therefore, complete and accurate cloud top height quality detection is realized.
Owner:NAT SATELLITE METEOROLOGICAL CENT

Coal seam gas content determination method

The invention discloses a coal seam gas content determination method. The method comprises the following steps: collecting a coal sample and placing the coal sample in a coal sample closed desorption container for closed normal-pressure desorption; opening the corresponding branches based on the change of the gas flow; working condition flow data and corrected gas flow data in the standard state are determined based on the laminar flow sensors on the branches and measurement data of the environment temperature sensor and the environment pressure sensor, and then the loss amount is calculated; the coal sample crushing device crushes the coal sample, and the corresponding branch is opened again based on the change of the gas flow until the desorption work is stopped; based on measurement data of each sensor, the coal sample mass is calculated to finally determine the coal seam gas content. The parameters such as the temperature, the pressure and the pressure difference of fluid flowing through the laminar flow sensor are measured in real time, the flow in a standard state is obtained after calculation and correction, and the measured gas flow and volume data are more scientific and accurate through temperature and pressure compensation.
Owner:CHINA COAL TECH & ENG GRP SHENYANG ENG CO

High-speed wind tunnel test flow field parameter uncertainty evaluation method and device and medium

The invention discloses a high-speed wind tunnel test flow field parameter uncertainty evaluation method and device and a medium, and relates to the field of aerospace craft wind tunnel tests. According to the method, by means of the characteristic that a general aircraft pneumatic database test needs more blowing train numbers, the influence of model condition replacement on flow field control and measurement result uncertainty is ignored, the test flow field conditions are analyzed and classified, the flow field parameter statistical uncertainty under the same test flow field condition for dozens of times can be conveniently obtained, the sample size is large, and the test efficiency is high. The reliability of the uncertainty result is high, the test content does not need to be increased, the economical efficiency is high, the process is simple and easy to implement, programming can be conveniently and rapidly achieved, and the assessment capability of the aerodynamic characteristic data accuracy of the aircraft can be effectively improved.
Owner:INST OF HIGH SPEED AERODYNAMICS OF CHINA AERODYNAMICS RES & DEV CENT

Bioelectrical impedance spectroscopy noninvasive blood glucose detection method

PendingCN119279574AImprove generalization abilityReduce the difference in bioelectrical impedance spectrumEnsemble learningSensorsHuman bodyElectric impedance spectroscopy
The invention relates to a bioelectrical impedance spectroscopy noninvasive blood glucose detection method, and relates to the technical field of noninvasive blood glucose detection. The method comprises the following steps: firstly, acquiring a bioelectrical impedance spectrum of the forearm of a human body by using a bioelectrical impedance sensor, then, reducing the difference of bioelectrical impedance between different subjects and sessions by using a proposed baseline correction method, and finally, expanding and balancing a sample based on an Optuna hyper-parameter optimization algorithm. Frequency and feature selection and regression model establishment are integrated into one evaluation index, and overall optimization of hyper-parameters is achieved. Specifically, the problems of small sample size and large blood glucose extreme value estimation deviation are simultaneously solved by using a data generation method based on K nearest neighbor in combination with a Clark error grid evaluation index and sample density weighting. Secondly, according to prior information of a data structure, using sparse group minimum absolute shrinkage with a weight proportion threshold value and a selection operator to simultaneously select frequency and characteristics which contribute to modeling to the greatest extent; and finally, an ensemble learning algorithm XGBoost is selected and used to establish a final model. The overall optimization method improves the robustness of the model and the accuracy of the estimation result.
Owner:GUILIN UNIV OF ELECTRONIC TECH

Joint optimization method for stacking sequence and grabbing posture of neutron measurement system

The invention provides a joint optimization method for a stacking sequence and a grabbing posture of a neutron measurement system, and relates to the field of data processing. According to the method, multi-source monitoring data of a neutron measurement system are collected, shadow voxels are constructed in combination with a three-dimensional space model, a neutron resonance shadow domain is formed through clustering superposition, and a dangerous concentration cloud cluster and a safe rarefied area are marked; mapping candidate stacking positions and postures according to sample mass distribution, geometric dimensions, centroid positions and irradiation sensitive characteristics, generating a stacking mode and determining a stacking sequence; performing path risk assessment on the grabbing postures, and selecting the grabbing postures of which the crossing risk meets a threshold value; superposing micro-random jitter on the path to suppress resonance phase locking; and abnormal events are monitored in real time during operation, and a risk domain is updated to execute hierarchical unloading control, so that the instantaneous impact and damage risk is reduced. According to the technical scheme, the probability of instantaneous impact risks caused by stacking overall micro-drift and grabbing track overlapping can be effectively reduced, and therefore the safety of the system is guaranteed.
Owner:CHINA STATE SHIPBUILDING CORP LTD RESEARCH INSTITUTE 719

Pesticide residue detection method and system for table vinegar production

The invention relates to the technical field of pesticide residue detection, in particular to a pesticide residue detection method and system for table vinegar production, and the method comprises the steps: obtaining a table vinegar raw material data set, carrying out adaptive clustering, and determining the sample size according to a preset proportion; the first detection end performs random sampling according to the sample size, generates a raw material danger identifier, uploads the raw material danger identifier to the first server, and predicts the vinegar production risk in combination with the monitoring data; and configuring sampling parameters, and controlling the second detection end to carry out vinegar detection. The technical problems that vinegar solution components interfere with pesticide residue detection and the pesticide residue detection accuracy is low are solved, monitoring from raw material purchasing as a starting point is achieved, a perfect raw material source tracing system is established through an informatization means, it is ensured that the source controls the pesticide residue risk, the pesticide residue detection accuracy of vinegar production is ensured, and the pesticide residue detection efficiency is improved. And an online real-time monitoring technology is adopted to capture and analyze pesticide residue information in raw materials or a processing process in real time, so that the technical effect of improving the detection efficiency is achieved.
Owner:JIANGSU HENGSHUN VINEGAR IND +2

Method for exploring performance goodness and generalization ability of multiple machine learning models in runoff sediment simulation

The invention provides a method for exploring performance goodness and generalization ability of various machine learning models in runoff sediment simulation. According to the method, the time sequence change trend of a river long-time sequence hydrological factor, a meteorological factor, a vegetation factor, an artificial activity factor and a runoff sediment transport change correlation factor is determined. Then identifying the correlation between each factor and the runoff sediment change, and selecting an important influence factor as an input factor of a machine learning model; the performance goodness of different machine learning models in simulation of runoff and sediment transport is evaluated, a sample data set is randomly divided to eliminate the influence of limited sample size and sample sequence on a simulation result, and each model is repeatedly trained and tested to determine the empirical risk, predictive ability and generalization ability of each model. And sorting the contribution degrees of the input factors in the runoff and sediment transport simulation process by applying a local interpretation method so as to reveal the main driving factors influencing the water and sediment change of the drainage basin.
Owner:BEIJING FORESTRY UNIVERSITY +1

Automated analysis device and analysis method therefor

PCT designated stage expiredWO2024241924A1Material analysisAutoanalysisProcess engineering
The purpose of the present invention is to provide an automated analysis device that enables high-precision sensing of abnormalities during aspiration even with small amounts of sample being used for analysis. To achieve the foregoing, this automated analysis device is provided with: a dispensing unit for aspirating a sample from a sample vessel accommodating the sample; a control unit for controlling the dispensing unit so as to aspirate a third predetermined amount of the sample, which is greater than a first predetermined amount that is the amount of the sample to be used for analysis, thereafter discharge a second predetermined amount of the sample to a predetermined place distinct from the sample vessel, and then discharge the first predetermined amount of the sample to a reaction vessel that is used for the analysis; a sensing unit for sensing a pressure waveform in the dispensing unit; and an assessment unit for assessing for air suction, which is a state in which there is insufficient aspiration of the sample by the dispensing unit, and / or clogging, which is a state in which the dispensing unit has been clogged with the sample, on the basis of a first waveform that is the pressure waveform of when the dispensing unit aspirated the sample.
Owner:HITACHI HIGH TECH CORP

Index transaction monitoring method and device, electronic equipment and storage medium

The invention discloses an index transaction monitoring method and device, electronic equipment and a storage medium, and relates to the technical field of transaction monitoring. The threshold range is dynamically generated based on the historical business data to adapt to the change of data distribution, the adaptive statistical test algorithm is selected in combination with the data sample size and the distribution characteristics to improve the accuracy of anomaly judgment, and personalized early warning output is performed according to the region and the business type of the auditing point. According to the technical scheme, data dynamic change and business scene diversity can be flexibly coped with, so that the technical problem that the reliability of an early warning result is reduced in the face of missing data, abnormal values and complex seasonal fluctuation, and then the real-time control and decision-making efficiency of an operator on a business operation state is influenced can be solved; the technical effects of improving the accuracy and reliability of business development index abnormity identification, reducing the false report and missing report rate, and enhancing the real-time control capability and decision-making efficiency of the operator for the business operation state are achieved.
Owner:CHINA MOBILE GRP HENAN CO LTD +1

Large-scale customization quality prediction method based on transfer learning

The invention discloses a large-scale customization quality prediction method based on transfer learning. The invention designs a large-scale customization quality prediction method for solving the problems that structures among variables are difficult to excavate and the quality of products with few samples is difficult to predict due to modular production, customized assembly and inconsistent number of different product samples in large-scale customization production. The method comprises the following steps: firstly, mining a causal relationship among source domain variables by utilizing a PC causal discovery method aiming at source domain data with a richer sample size, carrying out structure migration on a part of fixed structures from a source domain to a target domain by utilizing mechanism knowledge in order to obtain the causal relationship among target domain variables, and carrying out target domain few-sample amplification through a generative adversarial network; a causal relationship is complemented in a target domain after variable amplification through a PC causal discovery method, a target domain variable causal graph structure is obtained, then a heterogeneous causal graph attention network is utilized to carry out source domain and target domain model training, and target domain quality prediction is carried out through an MAML transfer learning method with MMD distance weighting. Aiming at the characteristics of large-scale customized production of customized products, the transfer learning-based quality prediction method suitable for large-scale customized production is formed.
Owner:CHINA JILIANG UNIV

Prediction method for port crude oil unloading speed regulation and control based on physical constraint

The invention relates to a port crude oil unloading speed regulation and control prediction method based on physical constraints, and belongs to the technical field of crude oil storage and transportation automation control, and the method comprises the following steps: S1, constructing a coupling equation of pipeline outlet pressure from an oil transfer arm to a reservoir area and storage tank flow speed; s2, constructing a physical constraint neural network model; s3, constructing a loss function formula of the physical constraint neural network model based on the coupling equation; and S4, performing adaptive training on the physical constraint neural network model based on historical data, and updating network parameters of the model. According to the method, the physical law of pipeline fluid movement is fused into the loss function in the form of constraint conditions in the model training stage, the degree of dependence of the model on the pure data sample size is remarkably reduced, and accurate regulation and control of the oil discharge rate under the complex working condition are achieved.
Owner:HAIBOTAI TECH (QINGDAO) CO LTD

Probability characterization method and system for design allowable value of thermoplastic composite material leading edge structure under small sample condition

PendingCN122024943AAchieve adaptive balanceTaking into account engineering practicalityChemical property predictionDesign optimisation/simulationProbability representationSmall sample
The invention belongs to the technical field of uncertainty probability characterization analysis, and discloses a thermoplastic composite material leading edge structure design allowable value probability characterization method and system under a small sample condition, and the method comprises the steps: defining a plurality of candidate probability distribution models; fitting each model based on the original sample data and calculating an AIC value and a BIC value; a dynamic weight factor alpha is calculated according to the sample size n, and then a hybrid information criterion HIC value is calculated; generating a plurality of sample sets through Bootstrap self-service sampling, recalculating the HIC value on each sample set, and counting the selected optimal frequency of each model; and determining an optimal probability distribution model according to the frequency, wherein the optimal probability distribution model is used for representing a design allowable value. According to the method, the dynamic weight factor alpha is introduced, AIC and BIC criteria are effectively unified, optimal balance between prediction precision and model complexity is achieved under the condition of small samples, and engineering practicability and robustness are remarkably improved.
Owner:AVIC XAC COMMERCIAL AIRCRAFT CO LTD

Carbon sink measurement system and method for multi-type urban green land ecosystem

The invention discloses a multi-type urban green land ecosystem-oriented carbon sink measurement system and method, belongs to the technical field of carbon sink measurement, and solves the problem of inaccurate evaluation result caused by non-uniform data distribution or insufficient sample size when a feature selection algorithm is adopted to extract feature factors in the existing method. The method comprises the following steps: collecting sequential urban green land carbon sink amount associated data, determining characteristic factors influencing the total carbon sink amount in a carbon sink preprocessing set based on a characteristic factor extraction model, analyzing a characteristic factor set by combining a carbon reserve method and the characteristic factor extraction model, and calculating the arbor and shrub biomass of the urban green land; according to the method, the feature factor extraction model is constructed by combining the GBR model and global sensitivity analysis, so that the feature factors influencing the total carbon sink amount in the carbon sink preprocessing set are accurately determined, the uncertainty in carbon sink estimation can be reduced, and the accuracy of an evaluation result can be ensured when data distribution is uneven or the sample size is insufficient.
Owner:ANHUI CONSTR ENG ECOLOGICAL TECH CO LTD

High-sensitivity drug impurity analysis and detection method and system

The invention discloses a high-sensitivity drug impurity analysis and detection method and system, and relates to the technical field of drug detection.The method comprises the steps that sampling is conducted based on drug batch information, and sample data is generated; constructing an impurity detection network, and generating an impurity signal; synchronizing the plurality of sample data to a chromatographic detection channel for analysis to generate a sample chromatogram; synchronizing a plurality of sample data to the spectrum detection channel for analysis, and generating a sample spectrogram; synchronizing the plurality of sample data to the mass spectrum detection channel for analysis, and generating a sample mass spectrum; evaluating the impurity signal, generating a sample impurity evaluation result, and extracting a drug impurity detection result. The technical problem of low medicine quality caused by low accuracy and incomplete detection of a detection method in the prior art is solved, and the technical effects of improving the comprehensiveness and accuracy of medicine detection and guaranteeing the medicine quality are achieved.
Owner:KEYPOINT PHARM TECH (QIDONG) CO LTD

Imported American cotton grading detection method and system based on neural network

The invention discloses an imported American cotton grading detection method and system based on a neural network, and relates to the technical field of cotton detection.By setting a threshold value and adopting a clustering algorithm to process uncertain samples, the recognition capability of abnormal and edge cases can be further enhanced, the high confidence of a detection result is ensured, and the detection accuracy is improved. According to the multi-level processing strategy, under the condition that the sample size is insufficient, accurate American cotton grading detection can still be achieved, in addition, the manual calibration link is introduced on the basis of the threshold value and the clustering algorithm, the problem that the sample size is insufficient, which is difficult to overcome in the prior art, is effectively solved, and the detection accuracy is improved. When a clustering algorithm cannot classify uncertain samples into known classifications, a manual calibration mode is adopted for processing at the moment, the accuracy and reliability of a detection result are ensured, meanwhile, the manually calibrated samples are fed back to a training set, continuous learning and self-optimization are effectively achieved, and the adaptability of a model to new situations is improved.
Owner:COMPREHENSIVE TECH CENT FOR INSPECTION & QUARANTINE OF ZHANGJIAGANG ENTRY EXIT INSPECTION & QUARANTINE BUREAU