87results about How to "Good prediction" patented technology

The embodiment of the invention provides a neural network system and method executed by a computer and used for analyzing a relational network diagram, and the neural network system comprises a feature extraction layer which is used for extracting feature vectors of nodes in the relational network diagram; a deep neural network used for carrying out first processing on the feature vectors to obtain first output; a diagram neural network used for combining the adjacency information of the relational network diagram and carrying out second processing on the feature vectors to obtain second output, wherein the adjacency information being used for representing a connection relation between nodes contained in the relation network diagram; and a fusion layer used for fusing the first output andthe second output and outputting a prediction result for the node based on a fusion result.

The invention discloses a wind power prediction method based on a modified particle swarm optimization BP neural network. The method includes the following steps: 1. encoding weight values and threshold values of a BP neural network as particles, and initializing the particles; 2. computing each particle fitness value with the difference between the result obtained from BP neural network training and an anticipated value as a fitness function; 3. comparing the fitness value of each particle and individual optimal particle to obtain a global optimal particle; 4. updating the speed and position of the particle; 5. determining whether the global particle meets termination conditions, if the global particle meets termination conditions, terminating the computing and outputting an optimal weight threshold value, and if the global particle does not meet termination conditions, back to step 2 and carrying out iterative operation; and 6. Using the optimal weight threshold value that is acquired by step 5 to connect an input layer, a hidden layer and an output layer of the BP neural network, and obtaining the result of wind power prediction on the basis of the result of the BP neural network. The method has fast convergence speed, high precision, and is not easily trapped to local extremum.

The invention discloses a short-term bus passenger flow prediction method based on depth learning and a passenger behavior mode, which comprises the following steps: 1, identifying and extracting characteristics of influencing factors of bus passenger flow; 2, reconstructing the data structure of bus passenger flow, reconstructing the input sample subdivision hour passenger flow xt into subdivision hour passenger flow matrix Xt, so that it can adapt to CNN and ConvLSTM models; 3. With historical passenger flow, the external and internal factors affecting bus passenger flow being used as inputdata, according to eight different dimensional data input schemes, namely, seven combined data input schemes considering internal influencing factors and one data input scheme without considering internal influencing factors, forecasting the bus passenger flow by using depth learning model, and obtaining the average relative error and absolute error of bus passenger flow forecasting through many experiments. The method simultaneously considers the external and internal factors of the bus passenger flow, and can not only predict the total bus passenger flow, but also predict the composition structure of the bus passenger flow.

The invention relates to an ultrashort combined predicting method for wind speed of a wind power plant. Firstly, wind speed time series of the wind power plant is acquired, and data pretreatment is carried out on the series to obtain input data of a prediction system; the input data is predicted respectively by a continuous prediction model, an ARMA (Auto Regressive Moving Average) prediction model and a wavelet-neural network prediction model, and three groups of prediction values are obtained through computation; and finally the three groups of prediction values are adopted to obtain the prediction value of the wind speed in the future 0-1 hour by a combined prediction method, and the last two groups of prediction values are adopted to obtain the prediction value of the wind speed in the future 1-4 hours by the combined prediction method. The combined prediction method is adopted, useful information of each single prediction method is fully utilized, the prediction precision of the wind speed in the future 4 hours of the wind power plant is improved, and a reference is provided for reasonably scheduling a power grid.

The invention discloses a real-time online individualized heat-exchanging intelligent temperature control system of mass concretes, and relates to the control on concrete water cooling in a high concrete dam casting procedure, wherein significance is provided for constructions of seamless dams. The system is characterized by comprising 1) installing a sensor in a newly casted mass concrete for measuring the temperature of the concrete in real time; 2) installing an interpolating digital temperature measuring device on water inlet / output pipes to measure the temperature of inlet and outlet water, and determining the average decreasing amplitude of the concrete temperature in real time according to a temperature difference of the inlet and outlet water; 3) determining a real-time flow according to the concrete temperature decreasing amplitude obtained in step 2); 4) installing an integrated flow and temperature control device integrated with temperature, flow and aperture controls on a water through pipe so as to realize real-time and online automatic collecting and feedback controls; 5) realizing the intelligent and individualized control on the mass concrete according to the temperature control information acquisition analysis of a temperature control acquisition instrument and the integrated flow and temperature control device and based on a temperature control gradient curve on time and space, so that an aim of casting the seamless dam is achieved.

A device is described for ascertaining state variables and / or parameters of a mathematical energy accumulator model, in particular a battery model, that describes the electrical properties of the energy accumulator on the basis of various state variables and parameters, the mathematical energy accumulator model having correction equations by which the state variables and / or the parameters are corrected and adjusted to the actual operating performance of the energy accumulator. A particularly rapid and accurate compensation of the energy accumulator model may be achieved if the error between a measured battery performance quantity and the battery performance quantity calculated by the energy accumulator models, a differentiated component of the error and an integrated component of the error are calculated, the individual components are each weighted using a weighting factor and the weighted error components are taken into account in correction of a state variable and / or a parameter.

The invention discloses an analog circuit fault prediction method based on ARMA (Autoregressive Moving Average), which comprises the following steps: extracting a plurality of characteristic quantities of a plurality of measuring points of an analog circuit to form a characteristic vector which can characterize fault information; utilizing an ARMA model to predict the characteristic vector to obtain a predicted characteristic vector; using the weighting Mahalanobis distance to calculate the distance between the obtained characteristic vector and a characteristic vector set in a circuit fault-free tolerance range; comparing the calculated distance and the maximum value of the Mahalanobis distance in the fault-free tolerance range, and converting the deviation degree of the calculated distance and the maximum value of the Mahalanobis distance to a fault occurring rate; and more intuitively monitoring the healthy state of the analog circuit. By experiment verification, the method can be used for better predicting the health state of the analog circuit, has a high fault detection rate, and can be used for the early monitoring of the analog circuit fault well.

The invention discloses a smart grid classification and fuzzy neural network based natural gas load prediction method and relates to the technical field of short-term load forecasting technology. The smart grid classification and fuzzy neural network based natural gas load prediction method comprises performing correlation analysis through Matlab to confirm an input variable; establishing a smart grid with a horizontal coordinate and a vertical coordinate to be the data and the average temperature through processed history data; selecting history data which is similar to a to-be-predicted date through the smart grid to train and predict a prediction model; performing de-noising processing through wavelets and training and predicting fuel gas loads through a fuzzy neural network with high adaptivity in combination with the complexity and multi-exterior-factor influences of the fuel gas load prediction process, wherein a process of modifying model parameters through errors is added to a training process and accordingly improvement of the final prediction accuracy is facilitated. The smart grid classification and fuzzy neural network based natural gas load prediction method can provide forceful reference for natural gas dispatching and confirms to material and technological foundation of a market development demand.

The invention discloses a risk model training method, a risk model training device, a risk identification method, a risk identification device, equipment and a medium. The risk model training method comprises the steps of: performing risk value labeling on historical travel data, and acquiring original training data; conducting companion analysis and port drift analysis on the original training data to obtain target training data; splitting the target training data according to preset time to obtain a training set and a test set; adopting a decision tree algorithm to train the target trainingdata in the training set, so as to obtain an original risk model; and adopting the test set to test the original risk model, so as to obtain a target risk model. The risk model training method effectively solves the problem that a current risk model has low identification efficiency and the accuracy rate of the current risk model is not high.

In a control loop for regulating a combustion process in a plant having a controlled system for converting material by way of the combustion, with at least one flame body being formed, the control loop having at least one observation device for imaging the flame body and further sensors to determine the state variables describing the state of the system in the plant, at least one regulator and / or a computer to evaluate the state variables and select suitable actions based on a process model, and adjustment devices for at least the supply of material and / or air that can be controlled by the actions, the process model provides specialized function approximators for various process dynamics, one of which function approximators is selected by a selector, and a regulator assigned to the selected function approximator is used to regulate the control loop.

The invention relates to a big data-based time-space confusion exposure degree assessment system and method. The system comprises a time-space data mining block, a multi-source heterogeneous data fusion module, a final variable selection module, a time-space generalized additive model building module, a re-sampling model module, a variation function time-space modeling module and a concentration estimation module; massive time-space data is mined; a relationship between multiple influence factors and pollutant concentration is established by adopting an accumulative nonlinear method; and through residual variation function fitting, spatial autocorrelation is considered, so that the prediction precision and effect are greatly improved.

The invention discloses a multi-dimensional electric energy meter on-site state test method based on a decision-making tree mass. The method comprises the following steps of building a decision-makingtree by using collected data related to an electric energy meter and provided by an electricity information collection system; forming next-level branches through classifying each decision-making tree; generating the completely grown primary decision-making tree; pruning and optimizing on the primary decision-making tree based on a cost complexity pruning algorithm; using an independent test setto assess the accuracy of the primary decision-making tree after pruning and optimizing; selecting the optimal decision-making tree from each set of secondary decision-making trees, and forming a decision-making tree abnormity diagnostic model by all the optimal decision-making trees, wherein a judgment result wins in a voting manner; and transmitting an operational rule of the decision-making tree abnormity diagnostic model to an electricity information collection system Hadoop big data processing cluster, thus implementing abnormity detection on the abnormal running state of the running electric energy meter.

The invention discloses an adaptive soft measurement prediction method based on a Bayesian network with a sliding window. According to the method, the advantages of the sliding window are fully exerted, the soft measurement model is updated by continuously adding new samples and deleting old samples, and the data closest to the sample to be predicted in the aspect of time are constantly selected to perform modeling. The prediction model is established by using the Bayesian network method in each window, the prior probability distribution of each node in the network is acquired through the parameter learning mode, and input information of the sample to be predicted acts as the evidence to be added to the established network to obtain posterior probability distribution of the node to be predicted so that the output mean and variance can be obtained and qualitative variable prediction can be completed. Compared with other present methods, accurate quality forecast can be given to the constantly changing industrial process so that the prediction value can be obtained and the corresponding prediction accuracy can be given, and the soft measurement prediction problem in case of missing of the data set can be greatly solved.

The present invention discloses an implicated crime principle and network topological structural feature based recognition method for drug-target interaction. The method comprises: firstly, according to human protein-protein interaction data and drug-target interaction data, constructing a drug-target interaction group network which comprises a protein-protein interaction sub-network, a drug-target interaction sub-network and a drug-drug relationship sub-network; according to information of a protein primary structure descriptor, a fingerprint feature of drug molecules and the reliability of interaction, weighting nodes and edges in the network; proposing a new network topological structural feature for characterizing a drug-target interaction pair based on an implicated crime principle and a graph theory; and finally, constructing a model by using a random forest algorithm and predicting a potential drug-target interaction effect in a proteome scale. The method does not require information of three-dimensional structures and the like of protein and drug molecules, is simpler, quicker and more accurate, and has high potential for application to the fields of new drug research and development, pathological study and the like.

The invention discloses an automatic debris flow wireless monitoring prevention and early warning device. The automatic debris flow wireless monitoring prevention and early warning device is characterized by comprising a field monitoring device and an indoor monitoring device, wherein the field monitoring device and the indoor monitoring device are in wireless communication connection. The field monitoring device comprises a laser displacement sensor (2), a rain gauge (3) and a strain gauge (4), wherein the laser displacement sensor (2), the rain gauge (3) and the strain gauge (4) are connected with a wireless transmission module (1). The laser displacement sensor (2) comprises a laser transmitter (5) and a laser receiver (6), wherein the laser transmitter (5) and the laser receiver (6) are arranged in an opposite mode. The wireless transmission module (1), the rain gauge (3) and the laser receiver (6) are arranged on a stable rock mass, the laser transmitter (5) is arranged on a sliding mass (10), and the strain gauge (4) is arranged in a region which debris flow easily flows through. According to the automatic debris flow wireless monitoring prevention and early warning device, field monitoring and indoor monitoring are combined, multiple prevention and early warning measures are taken, the prevention and early warning effect is remarkable, and the requirement for early warning of debris flow in a remote mountain area can be met.

The invention provides an unmanned aerial vehicle movement speed prediction method based on machine learning. The unmanned aerial vehicle movement speed prediction method comprises the following steps: S0: controlling an unmanned aerial vehicle to complete flight, in the period, acquiring an optical flow value of the unmanned aerial vehicle by using an optical flow meter, acquiring posture information and acceleration information by using a gyroscope and an accelerometer; S1: acquiring flying speed information according to the optical flow value and the posture; S2: training acquired data by using a Libsvm tool, thereby obtaining the relationship of the flying speed information and the posture information and the acceleration information; S3: when the unmanned aerial vehicle is used practically, predicting by using the Libsvm tool according to the relationship confirmed in the step S2 and the posture information and the acceleration speed acquired at present, thereby obtaining the movement speed of the unmanned aerial vehicle.

The invention discloses an inter-frame prediction method, a terminal devicet and a computer storage medium. The inter-frame prediction method comprises the following steps: constructing a candidate motion vector list of a current coding block; performing first prediction on the current coding block by using the candidate motion vector list to obtain a first prediction result, wherein the first prediction result comprises a plurality of first motion vectors and a first reference frame pointed by the first motion vectors; determining at least one prediction combination by using the first prediction result, wherein each prediction combination comprises two different first reference frames; carrying out symmetric vector residual prediction by using the prediction combination and the candidatemotion vector list corresponding to the prediction combination to obtain a second prediction result; and selecting the second prediction result with the minimum prediction cost as the symmetric motionvector residual prediction result of the current coding block. Through the inter-frame prediction method, the accuracy of the prediction value can be improved, the time redundancy is further removed,and the compression ratio of inter-frame coding is improved.

The invention discloses a mouth opening / closing state detection method based on deep learning. The method mainly comprises a data preprocessing portion, a feature extraction portion, a feature classification portion and an error calculation portion. The method is mainly characterized by, through full utilization of the capability in extracting high-level features of a depth convolution nerve network, extracting robust features capable of coping with irregular noise, larger illumination change and the cases of hostile attack and the like by sheltering the mouth key part, which occur frequently in the practical application scene; carrying out classification on the extracted features by utilizing a full connection layer; and adjusting parameters through error calculation and through a stochastic gradient descent method to reduce errors, and allowing the detection method to be able to have a capability of identifying mouth opening / closing state automatically. Besides, the method can ensure both the required computing resources and storage space do not generate great fluctuation due to change of resolution of an image to be detected. The method is convenient to operate, simple and easy to use, higher in precision and safer and more reliable.

The invention discloses a biological age step-by-step predication method based on a support vector machine, and relates to the biological age step-by-step predication method based on the support vector machine. The invention aims to solve the problems that a conventional biological age predication method is low in predication efficiency, low in accuracy, high in cost, and complex. According to the technical scheme, the biological age step-by-step predication method comprises: step I, preparing a biological age data set; step II, distinguishing a biological sample with known ages from a biological sample with unknown ages; step III, carrying out inter-group classifying; step IV, generating a corresponding support vector machine model; step V, establishing an optimal support vector machine model; step VI, establishing an optical characteristic sub set; step VII, obtaining the group type of an age group corresponding to the biological sample with known ages in the test set; step VIII, carrying out inter-group classifying; step IX, generating an inter-group classified support vector machine model; and step X, obtaining the exact age of a test set sample in certain age group. The biological age step-by-step predication method is applied to the biological age predication field.

The invention discloses a method for detecting the quality of Brassica campestris based on the near infrared spectrum, which comprises the steps of fitting collected near infrared spectrum data of a Brassica campestris sample and chemical values measured by a standard method, and optimally establishing a model by applying partial least squares (PLS); selecting an optimal spectrum pretreatment method, and measuring the quality of the model by comparing a determination coefficient (R2) and a root mean square error (RMSECV) so as to construct a high-quality quantitative analysis model of the nearinfrared spectrum of Brassica campestris. According to the method, the surface color, the mass loss rate, the hardness and the VC content of Brassica campestris can be rapidly and accurately predicted, the grade of Brassica campestris can be judged, thereby laying a foundation for rapid and nondestructive quality detection research of Brassica campestris, and having very high practicability and wide applicability.

The invention discloses a determination method for wheat root elongation toxicity of nickel ions under a hydroponic condition and application and belongs to the field of ecological heavy metal toxicity assessment. A control variable method is adopted, concentration of positive ions is controlled and changed separately, the inhibiting effects of five types of positive ions including K<+>, Na<+>, Ca<2+>, Mg<2+> and H<+> on nickel ion toxicity are calculated out step by step, meanwhile a MINTEQ program is utilized to calculate the amount of Ni<2+> and the amount of NiHCO3<+> under all conditions so as to calculate the toxic effects of two nickel forms, biological toxicity inhibition parameters of all the positive ions on Ni and toxic effect parameters of the two Ni forms are determined, the method for determining wheat root toxicity of Ni<2+> is established under the hydroponic condition, an obtained result can be used for prediction of wheat root toxicity of the Ni<2+> under the hydroponic condition, and a reference is made to soil heavy metal ecological risk assessment and soil contamination treatment.

The invention relates to an alternating current conductivity and dielectricity prediction method of a graphene-polymer porous nanocomposite material with consistent orientation based on an effective medium method. The prediction method comprises the following five steps of 1, testing the geometrical parameters and the electrical properties of a component material; 2, preparing a graphene-polymer porous nanocomposite material sample; 3, establishing an equivalent alternating current conductivity and dielectricity prediction model; 4, calculating and extracting the material parameters and obtaining a complete prediction model; and 5, obtaining a prediction curve and checking the prediction model. According to the method, the influence of the microstructures and / or the parameters such as theporosity, the graphene content, a slenderness ratio of graphene, a maximum included angle between the graphene, a percolation threshold and the like on the electrical properties of a product is mainlyconsidered, and the prediction model is re-established. After the model is checked, a prediction result is found to be closer to an experiment value.

The invention belongs to the field of signal processing, and particularly relates to an industrial alarm flooding prediction method based on an N-gram model. The industrial alarm flooding prediction method based on the N-gram model comprises the following steps that (1) a historical alarm flooding data set is acquired, alarm variables therein are counted, a discrimination degree of each alarm variable is calculated, and the alarm variables with 0 discrimination degree are eliminated; (2) sequences in the processed data set are compared with emerging sequences in similarity one by one, and thesequences are arranged from high to low according to similarity scores; (3) a time window is set to segment the reprocessed data set, the number of each data segment is counted, and the next possiblealarm variable and the corresponding probability are calculated by using a sample data set; (4) the probability of predicting the next alarm and a corresponding confidence interval are calculated through a Bayesian probability model; and (5) iterative operations are performed on the steps (3) and (4). The industrial alarm flooding prediction method based on the N-gram model solves the problem of inaccurate prediction of carrying out alarm flooding prediction at present.

The invention discloses a hydrocarbon source rock total organic carbon content prediction method considering a density factor. The method comprises the following steps: drawing the organic carbon content of a rock core of each well and a logging curve of the corresponding well on the same graph through software; segmenting the hydrocarbon source rock according to the maturity or age stratigraphictable of the hydrocarbon source rock; manually picking up the baseline value of the RD and the baseline value of the DT of each section; and according to the baseline value of the RD of each section,the baseline value of the DT of each section, the RD logging curve value corresponding to each depth and the DT logging curve value corresponding to each depth, solving the amplitude difference deltalogR of reverse superposition of the DT curve corresponding to the measurement point of the rock core of the multiple wells and the RD curve, and then predicting the total organic carbon content of the hydrocarbon source rock. According to the method, the tedious process that a traditional delta logR method needs to correspond to a maturity parameter chart is avoided, the influence of the compaction effect on the hydrocarbon source rock is considered, the application range of the traditional method is expanded, and the method has a good effect on the continental facies deep hydrocarbon sourcerock in China.

The invention discloses a renewable energy grid-connected consumption information multi-dimensional check method and equipment. The method comprises the follow steps: S1, a relationship model betweenelectric quantity and various meteorological factors is constructed by using a support vector machine algorithm to predict the power supply situation; and S2: an abnormal power consumption identification analysis model is constructed to identify the situation of abnormal power consumption based on the K nearest neighbor classification algorithm. Compared with the methods and the equipment in the prior art, the invention discloses the renewable energy grid-connected consumption information multi-dimensional check method and equipment, the support vector machine can be used for performing effective training and can be used for obtaining very good prediction accuracy. The training time of the classifier can be shortened and good performance in prediction accuracy can be shown.

Disclosed are systems and methods for rapid generation of simulations of a patient's spinal morphology that enable pre-operative viewing of a patient's condition and to assist surgeons in determining the best corrective procedure and with any of the selection, augmentation or manufacture of spinal devices based on the patient specific simulated condition. The simulation is generated by morphing a generic spine model with a three-dimensional curve representation of the patient's particular spinal morphology derived from existing images of the patient's condition.

The present invention relates to the technical field of laser welding, specifically to a prediction method for corrosion resistance of a laser welding plate, and a system thereof. The system comprises a database, a pre-processing module, a corrosion resistance prediction module and a post-processing module, wherein the pre-processing module reads base information and process parameters of welding parts and base materials required by a laser welding process from the database to provide initiate conditions for the follow-up process, the corresponding optimal PLS prediction model and corrosion resistance rate prediction values having a target relative error of less than or equal to 5% can be obtained with the corrosion resistance prediction module through simulation of different process states, and based on the corrosion resistance rate prediction values, reverse calculation is performed according to an inverse mapping principle to obtain the optimal welding process scheme corresponding to the minimum prediction value, wherein the optimal welding process scheme is the combination of the optimal process parameters.

The invention relates to the technical field of wind power prediction, in particular to a wind power output prediction method based on a historical predicted value. The method comprises the following specific steps: S1, building a BP neural network model, and carrying out the data quality analysis of historical measured data; s2, verifying BP neural network example indexes according to the data RMSE and nMAE; s3, after the index verification is passed, fragmenting the collected long-term historical data according to time, optimizing a neuron threshold by using a particle swarm algorithm, and performing network iteration neuron parameter correction by using a historical data fragmentation iteration process; and S4, after the parameter model is corrected, determining a predicted output time period according to an actual demand, carrying out time sequence series connection on historical data and a wind power prediction value of the wind power plant, taking a series connection result as input of the neural network model, and carrying out actual prediction value output and verification. The predicted value is closer to the actual value, and the requirements of current development are met.

The present invention discloses an oil-gas microorganism gene exploration method. The method comprises the following steps: respectively collecting samples of surface shallow layers above a known oilwell, a gas well and a dry well in an exploration area, performing high-throughput sequencing after extracting DNA, establishing a microbial community composition pattern diagram in the exploration area according to a sequencing result, respectively screening out characteristic microorganisms in surface soil above the oil / gas well in the exploration area according to the pattern diagram, and designing primers according to attribute characteristics and carrying out a fluorescent quantitative PCR detection on the samples in the whole exploration area to detect the number of the characteristic microorganisms. A contour line of the characteristic microorganisms obtained by the oil-gas microorganism gene exploration method has a high coincidence rate with the known well, has a good coincidencerate with a trap line, and can carefully describe an oil area.

The invention relates to a building method of an anthocyanin antioxidant activity three-dimensional quantitative structure-activity relationship model. An anthocyanin antioxidant with a known activity is taken as a study object, a technology of three-dimensional quantitative structure-activity relationship is used, an anthocyanin antioxidant three-dimensional quantitative structure-activity relationship model is built, an accurate structure-activity relationship model is built through a molecular force field method and a molecular similarity coefficient analytical method for analyzing technologies of molecular conformation optimization, parameter optimization and the like. The building method of the anthocyanin antioxidant activity three-dimensional quantitative structure-activity relationship model can quickly predict activity value of an unknown activity compound, and can reasonably interpret the relation between the size of the nthocyanin antioxidant activity and the structure characteristics of the nthocyanin antioxidant activity.