166 results about "Early prediction" patented technology

The invention discloses an edge cloud collaborative industrial equipment health management and predictive maintenance method, and the method comprises the steps: forming a closed-loop system through four deep coupling steps: edge adaptive fusion perception, cloud knowledge enhancement reasoning, edge cloud collaborative self-evolution prediction, and risk-driven maintenance decision. The edge end dynamically adjusts a sensor acquisition and feature fusion strategy according to the equipment health state, the cloud end performs causal reasoning and situation evaluation by using a physical-data mixed knowledge graph, and the edge cloud collaboratively optimizes a prediction model through federal element learning and bidirectional knowledge distillation; maintenance decision is based on multi-objective optimization, the actual effect is fed back to the perception and prediction link, the method achieves accurate assessment of the equipment health state, early fault prediction and maintenance intelligent decision, the availability of the equipment is remarkably improved, the maintenance cost is reduced, and core technical support is provided for intelligent manufacturing.

The invention discloses a pre-eclampsia prediction method and system based on machine learning and early pregnancy indexes. The pre-eclampsia prediction method comprises the following steps: acquiring early pregnancy laboratory index data and FMF model evaluation parameters of a to-be-predicted pregnant woman; based on a preset preeclampsia type, performing feature screening on the early pregnancy laboratory index data by using a Boruta algorithm to construct corresponding original features; inputting the original features into a trained first-level machine learning prediction model corresponding to the pre-eclampsia type to obtain a risk score; according to the risk score and an FMF model evaluation parameter, obtaining a joint feature; and inputting the joint features into a trained second-stage machine learning prediction model corresponding to the pre-eclampsia type to obtain a corresponding pre-eclampsia risk prediction result. According to the method, information of different sources and different physiological dimensions is deeply fused, analyzed and judged, and high-precision and low-cost early prediction of different pre-eclampsia types is realized.

The invention discloses a power load prediction and energy storage optimization control method and system, belongs to the technical field of power system energy storage optimization control, and aims to solve the problems that the reaction is slow, the charging and discharging mode is not optimized enough and the risk of countercurrent cannot be pre-judged in advance due to the fact that reverse power transmission to a power grid is prevented through detection of an anti-countercurrent electric meter in an existing anti-countercurrent method. Based on historical load data, environmental data and time features, a time sequence deep learning model fused with an attention mechanism is utilized to output a future multi-time-scale user net load prediction result in a rolling manner, and whether a grid-connected point power prediction value is smaller than a dynamic countercurrent threshold value or not is judged so as to identify a countercurrent risk. And if the risk exists, establishing a countercurrent avoidance optimization model, otherwise, establishing a conventional scheduling optimization model, performing rolling solution by adopting an efficient solution algorithm to obtain an optimal charging and discharging power instruction sequence, performing electric energy quality evaluation and correction, and then issuing the optimal charging and discharging power instruction sequence to the energy storage converter for execution, thereby realizing pre-judgment of the countercurrent risk in advance and optimization of energy storage charging and discharging behaviors.

The invention relates to the technical field of automobile safety, and provides an electric tail gate anti-pinch control method and system, which introduces a prediction + self-adaption intelligent decision normal form, on one hand, the motion information of a tail gate is obtained, and the time sequence data of an obstacle is combined to be input into an LSTM model to predict the track of the obstacle; a collision prediction conclusion is output to be matched with a graded speed reduction strategy to execute tail gate control, and through prediction in advance, the system response efficiency is effectively improved, and the clamping damage probability is reduced; and on the other hand, when a contact signal of the tail gate is detected, real-time contact force and real-time vehicle working condition information are obtained in real time, a dynamic anti-pinch threshold value is calculated and compared with the real-time contact force for analysis so as to execute tail gate control, the dynamic anti-pinch threshold value is dynamically adjusted according to the vehicle working condition and the tail gate state, scene adaptation is carried out, the anti-pinch recognition accuracy can be improved, and the safety of the vehicle is improved. The false triggering probability is reduced; therefore, the electric tail gate anti-pinch system which is faster in response, more intelligent and more reliable is obtained.

Methods and systems including computer programs encoded on computer storage media, for a method for detecting signals related to subjects participating in a clinical trial. In some implementations, a computer collects clinical data from multiple sources. The computer standardizes and redacts personally identifiable information and determines predictive features that correspond to characteristics of the data that correlate with efficacy and safety signals. The computer obtains training data and trains machine learning models to predict one or both of a predicted efficacy signal and a predicted safety signal for a subject. The computer receives clinical data for a particular clinical trial subject enrolled in an ongoing clinical trial and predicts one or more signals and receives a review of the signals through a user interface. The computer updates the predictive components of the system based on the review.

The invention discloses a power transmission line bird damage prediction method and device, electronic equipment and a storage medium. The method comprises the following steps: analyzing multi-source heterogeneous data based on a risk prediction model to obtain a target risk probability value of bird damage in a power transmission line area; the risk prediction model comprises a first model, a second model, a third model and a fourth model; the first model is used for predicting bird damage risks based on the meteorological data and the biological characteristic data; the second model is used for predicting bird damage risks based on the power transmission line data and the geographical environment data; the third model is used for predicting bird damage risks based on meteorological data; the fourth model is used for predicting bird damage risks based on bird damage defect data; and based on the target risk probability value and a preset risk grade range, predicting a bird damage risk grade of the power transmission line area, and carrying out risk early warning based on the bird damage risk grade. According to the invention, the problems of low risk assessment accuracy and insufficient early warning timeliness are solved, and early prediction and active early warning of bird damage risks are realized.

The invention provides a schizophrenia early recognition and risk prediction method and system based on multi-mode speech analysis, and belongs to the technical field of medical artificial intelligence. The method comprises the following steps: prospectively acquiring natural voice data of an individual in a clinical high-risk period, performing clinical follow-up visit for two years, and constructing a first Chinese schizophrenia clinical high-risk voice longitudinal data set; based on the data set, audio and text features in the voice are extracted, a layered multi-mode integrated model is constructed, semantic coherence and acoustic features are fused, and early prediction of the schizophrenia transformation risk is achieved; through feature importance analysis, key voice marks closely related to negative symptoms, thinking disorder and cognitive impairment are recognized, and non-invasive, accurate and interpretable early recognition is achieved. According to the method, the technical blank of early screening tools for schizophrenia in Chinese context is filled, and the accuracy and stability of early recognition are remarkably improved.

The invention relates to the technical field of grouting control, in particular to an intelligent grouting control method and system.The intelligent grouting control method comprises the steps that multi-parameter time sequence data in the grouting process is collected in real time, and standardization processing is conducted to construct feature vectors; inputting the feature vectors into an integrated learning model, and synchronously executing grouting effect prediction based on time sequence analysis and abnormal working condition identification based on data distribution analysis; when an abnormal working condition is recognized, a control strategy corresponding to the abnormity is executed preferentially; when abnormity is not recognized, the control parameters are dynamically optimized according to the predicted grouting effect; issuing the control strategy or the optimization control parameter to grouting equipment for execution, and collecting feedback data after execution; on the basis of feedback data, online incremental learning is carried out on the integrated learning model, model parameters are adaptively optimized, prediction can be carried out in advance, collected data and final control form a closed loop, meanwhile, the method can also adapt to complex working conditions, and the intelligent grouting control efficiency is improved.

A method for predicting early onset of mental health issues in work environments using artificial intelligence (AI) comprising: a) collecting data from two sources: (1) derived from employee survey an

The invention discloses a cross-battery few-sample early prediction method and system based on physical knowledge, and the method comprises the steps: obtaining the test data of a target battery through an accelerated aging experiment, and training a general life prediction model based on an iTransform architecture through a public data set; an improved dynamic time warping algorithm is provided to realize capacity ratio alignment of degradation tracks of the target battery and the reference battery, and a characteristic difference value sequence representing systematic deviation is extracted; further constructing a deviation prediction model of multi-task learning, and synchronously optimizing three sub-tasks of feature reconstruction, physical feature prediction and health state deviation estimation through an adaptive loss weight adjustment mechanism; and finally, carrying out personalized correction on the general prediction result through a double-model fusion strategy. According to the invention, the cross-battery degradation track prediction of the lithium titanate battery of the high-speed train is completed, the prediction accuracy and generalization are improved, and a reliable technical means is provided for the health management of the battery.

According to the helicopter thunder and lightning indirect effect simulation evaluation method provided by the invention, a simplified method of a complex aircraft model is provided, early prediction and protection design optimization of the thunder and lightning indirect effect of the aircraft are realized, the test cost is reduced, and the digital design and verification capability is improved.

The invention discloses an autism early prediction method and system based on multi-modal feature fusion and a bidirectional attention mechanism, and the method comprises the steps: carrying out the parallel extraction of an original video, carrying out the data enhancement through a segmentation recombination technology, and obtaining a training set, which comprises the enhanced behavior and physiological time sequence signals; double-flow deep feature learning is carried out on the training set by using a parallel flow frame, spatial-temporal features and time-frequency features are obtained, and the parallel flow frame comprises spatial-temporal feature extraction flow and time-frequency feature extraction flow; performing deep interaction and adaptive weighting on the spatial-temporal features and the time-frequency features by using a bidirectional cross attention mechanism to obtain fusion features; and inputting the fusion features into a classifier to obtain a final autism risk prediction probability. According to the method, the defects in the prior art are effectively overcome, and the accuracy and reliability of early prediction of autism are remarkably improved.

The invention discloses a preeclampsia biomarker and an application of the preeclampsia biomarker. The preeclampsia biomarker comprises at least one of a fusion gene PSPC1-MRPS31P2 and a fusion gene LINC00630-AL035494. The preeclampsia biomarker can be used for detecting preeclampsia. The preeclampsia biomarker provided by the invention can provide a non-invasive diagnosis scheme for clinical prenatal preeclampsia, and preeclampsia specific high-expression fusion genes (PSPC1-MRPS31P2 and LINC00630-AL035494) in peripheral blood of a pregnant woman are detected through real-time quantitative PCR (Polymerase Chain Reaction), so that preeclampsia early prediction, disease diagnosis and monitoring of the state of a patient in the illness period are realized; and in combination with the existing clinical diagnosis scheme, the accuracy of pre-eclampsia diagnosis is improved.

A wireless transmit / receive unit (WTRU) may be configured to perform measurements and measurement predictions in an idle state or an inactive state (e.g., low-power operation mode). The WTRU may perform measurements for the configured idle measurement durations. If the WTRU stays in the idle state or the inactive state for more than a validity duration after it has stopped performing measurements, the WTRU may perform measurement predictions. Upon transitioning to a CONNECTED state (full-power operation mode), the WTRU may report the actual measurements, if they are still valid, or otherwise, report the predicted measurements.

The application relates to application of phosphatidylinositol and phosphatidylglycerol as metabolic markers for predicting intracranial atherosclerosis progression, and belongs to the technical field of medical biological detection. In the scheme, single-sample MR analysis is carried out by combining RICAS cohort data with external database data, causal correlation between metabolites and ICAS is analyzed, and serum biomarkers phosphatidylinositol PI 18:0 / 22:6 and phosphatidylglycerol PG 18:1 / 18:2 for ICAS occurrence and development are screened out by combining characteristic metabolic changes of ICAS with causal correlation analysis. ICAS progression is collected through prospective follow-up of the RICAS cohort, the correlation between the serum biomarkers and the ICAS progression is established, and it is verified that the serum biomarkers have transformation application value and can be used as a new biomarker for early prediction of ICAS progression and derivation.

The application discloses an industrial valve fault prediction method based on multiple monitoring parameters, and relates to the technical field of industrial equipment fault prediction.The application realizes real-time collection of multi-dimensional operation data of a target valve, combines a fault prediction model to periodically output prediction data, accurately determines maintenance requirements based on comparison of a fault occurrence probability and a maintenance threshold, generates maintenance information and prompt information containing historical data traceability, realizes early prediction of fault risks, effectively avoids resource waste during a low fault occurrence period and intervention lag problems during a high fault occurrence period of traditional fixed cycle prediction, determines a plurality of intervention combinations and a prediction intervention coefficient through data analysis, dynamically determines whether intervention in the prediction process is needed based on a fault probability floating value of a continuous prediction cycle, determines an interval length of next fault prediction, realizes adaptive optimization of the prediction cycle, matches the fault prediction frequency with dynamic changes in risks, and improves the prediction response speed during a high risk stage.

A fault prediction method, device, equipment, medium and product are disclosed. The method comprises: obtaining a target data set, the target data set comprising: a container running index, an application log text, a listening event and a link tracking index; performing fault prediction based on the target data set to obtain a fault prediction result, the fault prediction result comprising: a fault probability, a fault type and a prediction time. Through the technical solution of the present application, real-time analysis of multi-dimensional indexes during container running can be performed to realize early prediction of faults.

The invention relates to a fault prediction system based on health management data of network switching equipment, and belongs to the field of fault prediction and health management. The system comprises a data acquisition module, a data preprocessing module, a fault mode tree module, a fault diagnosis module, a health assessment module, a fault prediction module and an early warning output module. According to the fault prediction method, hardware state and network state information of the network switching equipment are considered, real-time monitoring of equipment software and hardware operation conditions is achieved, the functions of automatic abnormal state diagnosis, system health assessment, fault prediction in advance and the like are provided, a necessary technical means is provided for health management of the network switching equipment, and the network switching equipment is protected. According to the invention, the transformation from post maintenance to condition-based maintenance and from periodic maintenance to maintenance based on operation state of the equipment is realized, and the maintenance support mode of the network switching equipment develops in a more effective direction.

The invention discloses a model for early prediction of acute kidney injury risk after a sepsis patient is transferred into ICU based on clinical variables, 1551 sepsis patients are included in the research, clinical variables of the patients in the ICU within 24 hours are collected, five most important clinical variables related to acute kidney injury are screened out, XGBoost is adopted to construct a prediction model, and the risk of acute kidney injury is predicted. The prediction model is good in performance, an SHAP method is introduced to generate a feature importance map and an individual prediction interpretation map, the contribution degree of each index to risk prediction of a specific patient can be visually displayed, and the cause of SA-AKI of the patient can be explained, so that a doctor can take targeted measures preventively.

The invention relates to the technical field of image recognition, in particular to a chicken flock health state inspection monitoring system based on image recognition, which comprises an individual behavior quantification module, a time sequence feature embedding module, a heterogeneous network construction module and a health risk assessment module. According to the method, each individual in the chicken flocks is endowed with a unique identity label, so that accurate quantification and continuous tracking of multi-dimensional behavior parameters of the individual can be realized, a dynamic behavior change file is established for each individual, subjectivity and discontinuity of traditional routing inspection are abandoned, and further, the accuracy of routing inspection is improved. According to the method, a complex social relation network is constructed according to the position proximity relation between individuals and the sharing condition of public resources such as water tanks and food troughs, and accurate evaluation and advanced prediction of individual health risks are achieved based on the dynamic behaviors of the individuals and comprehensive research and judgment of the states of the associated individuals in the network. Therefore, potential abnormal individuals are identified before disease diffusion, and the accuracy and timeliness of early warning are greatly improved.

The invention provides an early prediction method for breast cancer neoadjuvant therapy based on a pathological full-slide image, and belongs to the technical field of breast cancer auxiliary diagnosis and treatment. The method comprises the following steps: firstly, acquiring a breast cancer digital pathological image sample, and performing color normalization to obtain a preprocessed image; then, background removal and block cutting are carried out under the amplification factors of 10X and 40X, and features are extracted by using a pre-training standard model to obtain two groups of vectors; constructing a deep learning model, inputting two groups of vectors, performing multi-scale feature fusion analysis, and outputting a probability value; and finally, on the basis of semi-supervised multi-instance learning, dividing samples into a training set and a verification set in proportion, training the model, adjusting hyper-parameters according to the verification set, and retraining full samples to obtain a final model for prediction. The technical problems that tumor cell microscopic information cannot be mined through existing MRI image analysis, the manual labeling cost of a digital pathology full supervision method is high, and data size difference and tumor surrounding information mining are difficult to consider are solved.

This invention relates to a group of genes associated with gestational hypertension, a method for predicting gestational hypertension in early pregnancy using these genes, and related applications. Specifically, this invention obtains transcript start site feature values ​​from at least one of the following genes from the cfDNA sample of a pregnant woman: LOC124902572, EEF1A1P16, LOC105369767, GOLM2, CDRT7, MEIS3, and LOC105376108. These transcript start site feature values ​​are then input into a prediction model, and the output results are used to determine the risk of gestational hypertension. The number of sequences in the transcript start site region after correction based on the transcript start site feature values ​​is also considered. The prediction model obtained by this invention can effectively and robustly predict the risk of gestational hypertension (including preeclampsia) in early pregnancy, assisting doctors in early pregnancy intervention and reducing the risk of gestational hypertension.

The invention relates to the technical field of medical monitoring equipment, in particular to a newborn gastric motility intelligent monitoring and intervention system based on deep learning, which comprises a data acquisition module for acquiring electrogastrogram signals by adopting a flexible electrode array and acquiring body surface abdomen sound signals by adopting an acoustic sensor; the data processing module constructs an analysis framework comprising a topological feature extraction unit, a collaborative learning unit and a time sequence prediction unit based on a topological data analysis method, converts gastric motility signals into point cloud data in a high-dimensional feature space, calculates a persistent homology group, generates topological descriptors and constructs gastric motility state manifolds; gastric motility state recognition and prediction are realized through a deep neural network; according to the method, the internal structure characteristics of the gastric motility signal are effectively represented through a topological method, the detection rate of the gastric motility abnormality is remarkably increased, the gastric motility abnormality is predicted in advance, and a scientific basis is provided for clinical intervention.

This invention discloses a method for constructing an early prediction model of hydrocephalus secondary to intraventricular tumors based on the dynamic changes of inflammatory factors after surgery, belonging to the field of medical data processing and model construction. The method includes: acquiring the concentration values ​​of inflammatory factors and the hydrocephalus outcome of biological samples at multiple postoperative time points; calculating quantitative characteristic parameters characterizing the dynamic evolution of the inflammatory response, such as the dynamic change rate, cumulative load, and acceleration, based on the concentration values; and training a deep learning model using the quantitative characteristic parameters and clinical baseline data as input features, with the hydrocephalus outcome as the label. This invention achieves early quantitative warning of hydrocephalus risk, significantly improving the accuracy, sensitivity, and specificity of prediction. It solves the technical problem of existing technologies relying on post-operative monitoring methods such as imaging and lacking early biochemical warning capabilities.

The invention relates to the technical field of fault prediction of energy storage systems, and discloses a multi-level digital twinning fault prediction method for a high-voltage cascade energy storage system, and the method comprises the steps: firstly constructing digital twinning models of a battery monomer layer, an energy storage module layer, a bridge arm layer and a system layer for a hierarchical structure of the high-voltage cascade energy storage system; extracting fault feature parameters to form a multi-level fault feature set based on model output and measured data of each layer; performing weighted fusion on the multi-level features by using an attention mechanism to generate a comprehensive feature vector containing cross-level fault information; constructing a short-term, medium-and-long-term and fault propagation combined prediction model, and inputting a comprehensive feature vector to complete fault probability, evolution trend and propagation path prediction; and finally, determining a fault source hierarchy according to the residual anomaly of each hierarchy model, positioning a fault component through space-time clustering, and determining a fault source in combination with causal inference. According to the invention, through cross-level data fusion and fault propagation chain analysis, early prediction and accurate positioning of the fault are realized.

The invention relates to the technical field of individualized cognition, in particular to an individualized cognitive decline risk prediction method and system based on multi-mode data fusion. According to the method, multi-modal longitudinal evidence-based data is integrated on the basis of a group-level pathology law, a multi-scale causal link map capable of reflecting the complete period of cognitive decline is formed, pathology time axis alignment and individualized feature normalization are achieved, the dynamic pathology process of individual specificity can be captured, and the accuracy of the pathology process is improved. By fitting individual exclusive multi-scale causal parameters and constructing a time sequence cascade digital twinborn body, forward causal simulation can be executed on multiple time scales, a cognitive decline core driving path is directly revealed, and individual risk quantification and intervention targets are generated. Detection of early hidden pathological signals, dynamic modeling of individual heterogeneity and comprehensive analysis of a multi-channel pathological mechanism are achieved, the ultra-early prediction precision of cognitive decline and individualized intervention suitability are remarkably improved, and a full-period prediction system capable of being iteratively optimized continuously is formed.

The present invention discloses a set of novel epigenetic biomarkers for early prediction, treatment response, recurrence and prognosis monitoring of pancreatic cancer. Aberrant methylation of genes can be detected in tumor tissues and plasma samples from pancreatic cancer patients but not in normal healthy individual. The present disclosure also discloses primers and probes used herein.

This invention relates to a group of genes associated with preterm birth, methods for predicting preterm birth in early pregnancy using these genes, and their related applications. Specifically, this invention obtains transcript start site features from at least one of the following genes from the pregnant woman's cfDNA sample: TEX2, WNK1, LINC00970, LOC124900437, IGFL4, LOC105376571, LOC105376577, SERPINB10, LOC642554, and SLC26A7. These transcript start site features are then input into a prediction model, and the output results are used to determine the risk of preterm birth. The prediction model obtained by this invention has higher accuracy than existing models, providing more accurate clinical intervention guidance and reducing medical costs.

The invention relates to the technical field of foundation settlement prediction, in particular to a deep filling soil settlement control and prediction method under the action of dynamic compaction impact. Comprising the following steps: carrying out construction period settlement double-track prediction through fusion of an improved hierarchical summation method and FLAC3D numerical simulation based on a Duncan-Zhang constitutive model; constructing a post-construction settlement quantitative model, establishing a secondary consolidation settlement prediction formula considering the time effect, and performing post-construction settlement quantitative prediction; correction prediction is carried out, and a prediction result is corrected by adopting a practical correction coefficient matrix for dynamic compaction over-consolidated soil settlement calculation; four-in-one closed-loop control is adopted, a closed loop from prediction, control to verification is formed, and construction parameters are dynamically adjusted; through the mode, under the action of dynamic compaction impact, the settlement prediction precision of the deep and thick soil filling foundation is high, a complete technical closed loop from early-stage prediction, middle-stage construction control to later-stage monitoring verification is formed, and comprehensive and real-time prediction can be achieved.

The invention belongs to the technical field of automatic control, and discloses an anti-deviation positioning device for a transplanting mechanism, which comprises a data acquisition unit, a deviation prediction module, a self-adaptive PID control module and a deviation early warning module, the problems that in existing transplanting mechanism positioning control, PID parameter self-adaptive capacity is poor, neural network error modeling is insufficient, deviation risk prejudgment lags behind and the like are solved, and high-precision dynamic positioning and deviation risk active early warning in the transplanting process are achieved. According to the device, a PID-timing attention fusion control framework is innovatively introduced, the real-time adjustment characteristic of a traditional PID is combined with the multi-dimensional error prediction capability of a timing attention neural network, and real-time deviation correction and pre-judgment of positioning deviation are achieved through a double-branch cooperation mechanism; on the deviation prediction level, a multi-scale time sequence convolution-gating cycle unit (TCN-GRU) fusion model is constructed, multi-dimensional time sequence feature extraction is carried out on motion parameters and environmental perception data of the transplanting mechanism, and positioning deviation trends under different working conditions are accurately predicted.