504 results about "Artificial neural network model" patented technology

The invention discloses the technical field of wireless communication and wireless network positioning, and in particular relates to an improved positioning method of an indoor fingerprint based on a clustering neural network. According to the technical scheme, the positioning method is characterized by comprising the following steps of: an offline phase: constructing a fingerprint database by fingerprint information collected from a reference point, sorting fingerprints in the fingerprint database by utilizing a clustering algorithm, and training the fingerprint and position information of each reference point by utilizing a artificial neural network model to obtain an optimized network model; and an online phase: carrying out cluster matching on the collected real-time fingerprint information and a cluster center in the fingerprint database to determine a primary positioning area, and taking the real-time fingerprint information in the primary positioning area as an input end of the neural network model of the reference point to acquire final accurate position estimation. The method has the advantages that low calculation and storage cost for the clustering artificial neural network fingerprint positioning method can be guaranteed, the positioning accuracy of the clustering artificial neural network fingerprint positioning method can be improved, and accurate positioning information is provided for users.

The invention relates to a method for predicting the dynamic mechanical property of a material based on a BP artificial neural network, aims at achieving the prediction of the dynamic mechanical property of the material through the BP artificial neural network, and belongs to the testing field of the dynamic mechanical property of the material. The principle of the method comprises the steps: collecting stress-strain data through employing a high-speed tensile test method, and obtaining a training sample set after normalization preprocessing; building a BP artificial neural network model through designing an input layer, a hidden layer and an output layer, and selecting a proper transfer function, a training function, and a learning function; carrying out the iterative training of the BP artificial neural network through employing the training sample set, and obtaining an optimal prediction network. The above prediction method can be used for the prediction of the dynamic mechanical property of the material, can achieve the quick prediction of a flow curve of the material at different strain rates in a short time, and can provide enough sample data for automobile safety simulation.

A method is described for improving the prediction accuracy and generalization performance of artificial neural network models in presence of input-output example data containing instrumental noise and/or measurement errors, the presence of noise and/or errors in the input-output example data used for training the network models create difficulties in learning accurately the nonlinear relationships existing between the inputs and the outputs, to effectively learn the noisy relationships, the methodology envisages creation of a large-sized noise-superimposed sample input-output dataset using computer simulations, here, a specific amount of Gaussian noise is added to each input/output variable in the example set and the enlarged sample data set created thereby is used as the training set for constructing the artificial neural network model, the amount of noise to be added is specific to an input/output variable and its optimal value is determined using a stochastic search and optimization technique, namely, genetic algorithms, the network trained on the noise-superimposed enlarged training set shows significant improvements in its prediction accuracy and generalization performance, the invented methodology is illustrated by its successful application to the example data comprising instrumental errors and/or measurement noise from an industrial polymerization reactor and a continuous stirred tank reactor (CSTR).

An artificial neural network compression coding device comprises a memory interface unit, an instruction buffer memory, a controller unit and an operation unit, wherein the operation unit is used for performing corresponding operation on data from the memory interface unit according to the instruction of the controller unit. The operation unit mainly performs three steps of: 1, multiplexing an input neuron with weight data; 2, performing an addition tree operation for adding the weighted output neurons after the first step through an addition tree or obtaining a biased output neurons through adding the output neurons and a bias; and 3, executing a function activation operation, and obtaining a final output neuron. The invention further provides an artificial neural network compression coding method. The artificial neural network compression coding device and the artificial neural network compression coding method have advantages of effectively reducing size of an artificial neural network model, improving data processing speed of the artificial neural network, effectively reducing power consumption and improving resource utilization rate.

The invention discloses a robust control method of burden error of pitch concrete stirring device, which comprises first building an input/output artificial neutral network model of three-layer structure, while the input layer has one neuron, the middle layer has five neurons and the output layer has one neuron, processing dynamic prediction and learning. The invention builds the non-linear model of flux and fly ash amount to be compared with a linear model of fixed parameters via continuous learning and adjustment, to adapt the change of external parameters, thereby effectively controlling the burden error of pitch mixture stirring device, reaching +-2% burden error as maximum and improving the burden accuracy of pitch mixture stirring device.

A system and method for modeling technology to predict accurately water-oil relative permeability uses a type of artificial neural network (ANN) known as a Generalized Regression Neural Network (GRNN) The ANN models of relative permeability are developed using experimental data from waterflood core test samples collected from carbonate reservoirs of Arabian oil fields Three groups of data sets are used for training, verification, and testing the ANN models Analysis of the results of the testing data set show excellent correlation with the experimental data of relative permeability, and error analyses show these ANN models outperform all published correlations

The invention discloses an urban road traffic jam judging method based on vehicle GPS data, and relates to an urban road traffic jam judging method. The problem that an application range of a traffic jam judging method depending detection equipment data is relatively large in limitation because conventional traffic information detection equipment is adopted by an existing urban road traffic jam judging method is solved. The urban road traffic jam judging method comprises the following steps: constructing an urban road link travel time prediction model based on an artificial neural network model; calculating link travel time data of a current moment according to a position vector, a link number vector, a time stamp vector and a speed vector of the current moment according to a vehicle GPS by using the urban road link travel time prediction model; further calculating a link traffic flow velocity and a link traffic flow density based on the link travel time data; with data of the link traffic flow velocity and the link traffic flow density as input conditions, judging a road traffic jam state. According to the urban road traffic jam judging method, the traffic jam state can be rapidly and accurately judged according to the GPS data of the current moment.

The invention provides a bus arrival time estimation method and device. The method includes dividing each bus running line into a plurality of running portions, and establishing corresponded artificial neural network models respectively; training the running portions to acquire corresponded parameter values of the corresponded artificial neural network models; determining all target running portions between the current bus position and a target station, acquiring current values of the current influence factors corresponding to each target running portion, inputting the corresponded artificial neural network models, acquiring output values of the artificial neural network models corresponded to the target running portions, and acquiring times for the bus arriving the target station according to the output values. The device comprises a dividing module, a model establishing module, a first artificial neural network model processing module, a current state processing module, a second artificial neural network model processing module and a calculating module. By the aid of the method and device, the bus arrival time can be calculated more accurately.

The embodiment of the invention provides a method for applying seismic multiattribute parameters to predicting the coal seam thickness. The method comprises: a suitable time window is selected in a three-dimensional offset data body, seismic attribute data of amplitude, frequency, and instantaneity and the like are extracted from the time window, and a seismic attribute database is established; a correlated analysis is executed on seismic attributes and coal seam thicknesses and cross-correlation analyses are further executed on the seismic attributes, so that a plurality of seismic attributes that are most meaningful are optimized as basic parameters of a coal seam thickness prediction model; with combination of known boring data, a multicomponent polynomial regression model and a BP artificial neural network model of between all the seismic attributes and the coal seam thicknesses are established by utilizing a multicomponent polynomial regression method and a BP artificial neural network method; and the models are utilized to predict coal seam thicknesses. According to the method provided in the embodiment of the invention, because multiattribute parameters are considered, obtained calculating models are perfect and realistic; an effect for prediction of the coal seam thickness is good; and credibility and accuracy are high.

The invention discloses an ARM temperature and humidity self-correction based electromagnetic radiation measuring device and a measuring method. The ARM temperature and humidity self-correction based electromagnetic radiation measuring device comprises an electromagnetic radiation sampling module, a signal processing module, a temperature and humidity digital sensor, an embedded type micro-processor module, a memorizer and a display module; the embedded type micro-processor module is respectively connected with the signal processing module, the temperature and humidity digital sensor, the memorizer and the display module; the signal processing module is connected with the electromagnetic radiation sampling module. According to the ARM temperature and humidity self-correction based electromagnetic radiation measuring device, environment temperature and humidity influences to a measuring result are considered meanwhile the environmental electromagnetic radiation intensity is measured, the measuring result is corrected in real time through an artificial neural network model, the measuring result authenticity and accuracy is guaranteed, environmental limit to utilization conditions of instruments is effectively widened, measurement can be performed under the severe environment, and the device automatically stops working to avoid instrument damage when the environment temperature and humidity exceeds a threshold value.

A method for applying a bokeh effect on an image at a user terminal is provided. The method for applying a bokeh effect may include: receiving an image and inputting the received image to an input layer of a first artificial neural network model to generate a depth map indicating depth information of pixels in the image; and applying the bokeh effect on the pixels in the image based on the depth map indicating the depth information of the pixels in the image. The first artificial neural network model may be generated by receiving a plurality of reference images to the input layer and performing machine learning to infer the depth information included in the plurality of reference image.

The invention provides an optimization method with integrated circuit mask design. The optimization method comprises the steps of S1, providing a whole chip design layout of an integrated circuit, and randomly grasping multiple design layout small areas in the whole chip design layout; S2, conducting mask optimization which is based on pixels on the chosen design layout small areas, and outputting a pixel grey-scale map of mask layout of each design layout small area; S3, utilizing the design layout small area mask pixel grey-scale maps obtained in the step S2 and small area design layouts corresponding to the design layout small area mask pixel grey-scale maps, and establishing a BP artificial neural network model; S4, sending the whole chip design layout into the BP artificial neural network model established in the step S3 to obtain a mask design layout grey-scale map of the whole chip design layout. The invention further provides a medium of a computer program used for storing the integrated circuit mask design.

The invention discloses a double-channel pump optimization method based on a multi-objective genetic algorithm. Turbulence calculation and an optimization algorithm are combined in the method and through the optimization algorithm, an optimum solution of structure parameters of a double-channel pump is found, so that the method can provide reliable design basis for workers lacking of rich design experiences and the quality of the design is improved. The training sample of an artificial neural network model in an optimization model is obtained through CFD analysis; a large number of temporary pump hydraulic performances generated in the optimization process are obtained through approximation model prediction; and therefore, computational accuracy is guaranteed, optimization-finding process is greatly accelerated, and time is shortened.

An automatic recognition and classification of ore and mineral images are disclosed in the invention. the invention utilizes the computer vision technology and the depth convolution neural network theory, Based on the big data platform Tensorflow, the convolution artificial neural network model is established, and the image data input model is trained according to the microscopic photographs of yellow iron ore from Jiapigou Gold Mine in Jilin Province, so as to realize the automatic recognition and classification of different ore minerals in the microscopic photographs of yellow iron ore. Theinvention can assist geologists to identify and classify the microscopic photographs of ore minerals and improve the working efficiency of geologists.

The invention discloses a gray generalized regression neural network-based small sample software reliability prediction method. The method comprises the following steps of: first, respectively emulating and expanding failure time data and test coverage rate data in collected small sample software reliability test data by using an improved Bootstrap method to form expanded reliability data which has the same failure statistical rule as small sample reliability data; then, obtaining a three-dimensional curve of the failure time, the test coverage rate and the unreliability of the expanded reliability data; next, establishing a gray generalized regression neural network; later on, training the gray generalized regression neural network by adopting the expanded reliability data and establishing a small sample software reliability prediction model; and finally, predicting by using the model to obtain software reliability prediction information. The method avoids solving a complex multivariate likelihood equation, and solves the problem that an available prediction model can be obtained only by training a large number of models in artificial neural network modeling in software reliability prediction.

The invention provides a multifunctional testing system and method for natural gas hydrate. According to the invention, the multifunctional testing system and method can highly efficiently determine the generation mass, structure type, thermodynamic conditions and generation rate of the natural gas hydrate by utilizing the variable quantities of reactants participating in a reaction and environmental conditions, determine the advantages and disadvantages of a natural gas hydrate inhibitor and the like while producing the natural gas hydrate; and the testing method is a comprehensive testing and analyzing means, provides a simple and feasible manner for further research on the structural characteristics and changing rules of the natural gas hydrate and is of important significance to understanding of the formation mechanism, microscopic dynamics, phase transition and the like of the natural gas hydrate. In particular, an artificial neural network model is established on the basis of a training database formed by known data, so accurate determination of the structure types of the natural gas hydrate is realized, and a research direction is provided for rapid determination of the structure types of the natural gas hydrate.

The invention provides a holographic early warning method of a mine dynamic disaster. A modal parameter set directly influencing occurrence of the mine dynamic disaster is used as an input variable set, the probability of occurrence of rock burst and the probability of occurrence of coal and gas outburst are used as output variables, and an artificial neural network model is built; samples are added to a learning sample database, and a holographic modal early warning device of the mine dynamic disaster is obtained; when early warning is conducted on any area of a mine, relevant information obtained in a mine experimental test and safe production is converted into input parameters of an early warning model by means of a holographic data mining converter, and then an early warning result isgiven by means of the holographic modal early warning device. By means of the holographic early warning method, holographic modal online prediction and early warning of the mine dynamic disaster can be achieved, and the system can achieve early warning separately based on the probability of occurrence of the rock burst accident and the probability of occurrence of the coal and gas outburst accident.

The invention relates to a disease data processing method and device, electronic equipment and a computer readable medium, and relates to the field of medical big data processing. The method comprisesthe steps: obtaining disease data, wherein the disease data comprise at least one disease symptom tag; Performing word segmentation processing on the disease data to generate a vocabulary set; Constructing a symptom set through the vocabulary set, wherein the symptom set comprises at least one disease symptom tag; And inputting the symptom set into a diagnosis model to obtain a disease classification identifier, wherein the diagnosis model is an artificial neural network model. According to the disease data processing method and device, the electronic equipment and the computer readable medium, the disease prediction accuracy can be improved, and a better aid decision is made for diagnosis of clinicians.

The invention relates to a detection method and a detection device for storage battery surplus capacity. According to the detection method and the detection device, a plurality of capacitance predicted values and a plurality of surplus capacity predicted values of a storage battery to be detected are obtained by utilizing an artificial neural network model library and measured values of influencing factors of the surplus capacity, the plurality of capacitance predicted values are subjected to weighted fitting, a capacitance measured value is regarded as a target value, a capacitance fitted value is optimized by adopting a particle swarm algorithm so as to obtain an optimized weight coefficient, and the plurality of surplus capacity predicted values are subjected to weighted fitting by utilizing the optimized weight coefficient, so as to obtain a surplus capacity value of the storage battery to be detected. The artificial neural network model library is used in the detection method and the detection device, the plurality of accurate surplus capacity predicted values can be generated, the optimized weight coefficient is obtained by adopting a particle swarm algorithm after the plurality of surplus capacity predicted values are subjected to weighted fitting, so that the plurality of surplus capacity predicted values after weighted fitting are more close to an actual surplus capacity value of the storage battery to be detected, and the detection precision of the surplus capacity value is further improved.