31 results about "Statistical learning theory" patented technology

The present invention is a method for extracting electromagnetic parameters of artificial electromagnetic material based on support vector machine (SVM). The invention relates to a new method for researching electromagnetic parameter measurement, capable of testing shield-hand material and artificial electromagnetic material having a periodic structure, and the testing result precision is high andthe production of testing samples is simple. The support vector machine (SVM) method is based on a VC-dimension theory of a statistical learning theory and a structure risk minimum principle, seeks an optimum compromise between the complexity and the learning capacity of a model based on limited sample information so as to obtain best popularization capability, and is widely applied to statistical classification and regression analysis. According to the invention, transmission and reflection coefficients of material to be tested are calculated by numerical computation methods FEM and FDTD ofelectromagnetism, and the corresponding computed result is used as training sequences to train the the support vector machine. When the support vector machine is trained fully, it is capable of calculating equivalent dielectric constant and equivalent magnetic conductance of the material to be tested by inputting testing values of the transmission and reflection coefficients.

The invention relates to a method for analyzing the service state of numerical control equipment, belonging to the technology for monitoring the service state of major equipment and forecasting the service life of the major equipment. The method comprises the following steps: firstly, recognizing the service states of a plurality of the major parts of the numerical control equipment through the characteristic physical quantity acquired by a multisensor; then, forecasting the service state of the complete machine through a classification model of a support vector machine established by the statistical learning theory; and calculating the residual service lives of the major parts and the complete machine through a 'hidden semi-Markov' random process model. The method of the invention not only can be used for recognizing the current running states of the parts but also can be used for forecasting the residual service life of the parts. The current running state and the residual service life of the complete machine are obtained by the classification forecasting method of the support vector machine according to the operation result of each part. The invention provides a new method for decision support of preventive maintenance.

The invention discloses an effluent index online soft measurement prediction method in the urban sewage treatment process. According to the method, firstly an instant learning model structure with correlation as the principle and the recursive partial least-square method as a local model is adopted, and meanwhile, description on uncertainty of the model is performed on the basis of the statistical learning theory. The model can be used for fully reflecting the characteristics such as uncertainty and large lag of the sewage treatment process, the explanatory performance of a prediction result is higher, controllers can adjust process variables such as the aeration rate and the backflow amount of an aeration tank in time, material balance and bacterial population balance are kept, and efficient removal of organic matter is achieved fully.

In the system, two matching results of human face and hand shape can be obtained through image preprocesisng and pattern identifying technique separately after image of the human face and the hand shape is collected. The two results are normalized into the same range by utilizing standard normalization method and the decision blending for two obtained results are carried on by utilizing support vector computer method in statistic theory so that the final authentication result can be obtained as the judgement is worked out. The identification result of hand shape has realy a high accuracy and rubustness owing to that the degree of freedom for collecting hand shape image is increased in its identification algorithm.

The invention relates to a machine learning method based on statistical learning theory. In order to solve the problems on large-scale sorting and the solution optimization in practical realization of an SVM (support vector machine), and realizes control on time price and space price of calculation, the technical scheme adopted by the invention is a method for realizing support vector machine by adoption of MPI programming and OpenMP programming. According to the concept of sorting algorithm of SVM (support vector machine), the method is realizes as follows: serial program codes are complied by C++ and associated statements and functions of OpenMP and MPI based on the serial codes are added to realize parallelization. The method provided by the invention includes the following detailed steps: 1) determining functions of parts of the algorithm; communicating among the algorithm modules by MPI programming to transfer data and realize synchronization; and 2) adding compiling guidance statements by Open MP in the sub-modules of the algorithm, wherein a compiler automatically conducts thread-level parallel realization of the codes in the parallel area included in the compiling guidance statements. The method provided by the invention is mainly applied to the machine learning.

A method for activating at least one safety device that has a first step of acquiring at least two features from at least one signal of a crash sensor system in order to form a feature vector from the features acquired. In a second method step, the formed feature vector is subsequently classified with the aid of a classifier based on the statistical learning theory in order to classify the feature vector in one of at least three possible feature classes. As a third method step, the safety devices are activated in accordance with an activation instruction for the feature class in which the feature vector was classified.

The invention relates to a method for extracting and recognizing facial features. The invention refers to a method for extracting and recognizing a two-dimensional front facial image. The specific recognizing methods are as follows: (1) a method based on a feature face: projecting an original image to a feature space, obtaining a series of dimension-reducing images and taking a main element thereof to express a face; (2) a method based on a neural network: utilizing a system formed by linking a great mass of simple processing units; (3), a method based on a support vector machine: building an ability of precisely studying a specific training sample and correctly recognizing any samples on the basis of vapnik-chervonenkis (VC) dimensional theory of the statistical learning theory and the minimum principle of the structural risk. The method for extracting and recognizing facial features can effectively reduce the error rate of the feature face part analysis by adopting the method based on the feature face; the method based on the neutral network can get good effect by applying to the face detection and the front face recognition; and therefore, the problem of complex mode recognition and action control can be solved.

The invention discloses a method for measuring and calculating water requirement of regional crops, comprising the following steps of: selecting the measuring and calculating area, invoking DEM, crop planting structure, soil type, weather factors and water requirement data of the crops about the measuring and calculating from a system database; interpolating the terrain factor data and the weather factor data to obtain a grid database set of regional impact factors; extracting comprehensive factor date having an impact on the crop water requirement from each terrain factor data, soil type data and weather factor data; establishing nonlinear mapping relationship between the comprehensive factor date and the crop water requirement to acquire the water requirement data of the different crops; the invention can use diverse space processing methods aiming at diverse impact factors, sufficiently takes account of the space variability of each impact factor, and through the change process of original data and the application of statistic theories, the invention can efficiently reduce the operation workload, improve the data analysis efficiency, and enhance the measuring and calculating precision for the water requirement of regional crops on condition of complicated terrain and sparse site.

The invention discloses an industrial process fault diagnosis method based on information fusion. A KPCA data processing method can be used for dealing with the nonlinear problem of data. Two machinelearning methods of SVM based on a statistical learning theory and SOM based on an artificial neural network are used to realize decision of a fault type, a D-S evidence theory is used to carry out information fusion on a decision result, and an obtained final diagnosis result has higher fault diagnosis precision. According to the invention, a plurality of methods are fused to replace an originalsingle method to process fault diagnosis of a complex industrial process, so that the multi-fault type diagnosis capability and the fault diagnosis precision are improved.

The invention relates to a wind speed forecasting method and device based on G-L mixed noise characteristic kernel ridge regression technology. The method comprises the following steps: 1) obtaining wind speed data set D1; using the Bayesian principle for the loss function of Gaussian-Laplace mixed noise characteristic; 2) through the use of the theories of statistical learning and optimization and in combination with the loss function obtained in step 1), establishing the original problem of the kernel ridge regression model based on the Gauss-Laplace mixed noise; deducing and solving the dual problem of the kernel ridge regression model; 3) determining the optimal parameters of the dual problem of the kernel ridge regression model; selecting the kernel function; constructing the decision function of the kernel ridge regression model; and 4) constructing the wind speed forecasting model of the kernel ridge regression model; and using this forecasting mode to forecast and analyze the wind speed value. The device of the invention includes a loss function obtaining module, a dual problem solving module, a decision function constructing module and a wind speed forecasting module. The method and invention meet practical application in wind power generation, agricultural production, and etc. which are demanding in terms of wind speed forecasting accuracy.

The invention discloses an XLPE (Cross Linked Polyethylene) cable partial discharging signal estimation method based on a statistical learning theory. The XLPE cable partial discharging signal estimation method comprises the following steps: inputting an acquired XLPE cable partial discharging signal; representing the XLPE cable partial discharging signal by linear combination of a primary function; carrying out wavelet decomposition on the XLPE cable partial discharging signal to obtain a new wavelet coefficient sequence from big to small according to decomposition coefficient energy; enabling VCdimensionh to be equal to 1, 2,...L in sequence, and keeping former (h-1) wavelet decomposition coefficients from the new wavelet decomposition wavelet decomposition; zeroing residual decomposition coefficients and calculating corresponding structure risks; finding out a VCdimensionh0 which enables {Rstr(h)} to be minimum; and outputting an optimal estimated signal of XLPE cable partial discharging. The XLPE cable partial discharging signal estimation method based on the statistical learning theory has the beneficial effects that the matching degree of a treated signal waveform and a real signal waveform is high, a discharging signal component can be kept better, and the smooth effect on background noises is very remarkable.

The invention provides a stem cell activity detection system based on a svm algorithm. The system comprises a subset generation unit, a subset evaluation unit, a stop condition unit and a result verification unit. A support vector machine is widely applied to each filed of machine learning. Correlation problems of classification and mode identification can be solved, and problems of regression, fitting, density estimation and the like can be effectively processed. Compared to other existing machine learning algorithms, the SVM has a solid theory foundation and research and application are wide. The SVM is a small sample learning algorithm based on a statistics learning theory, structure risk minimization is taken as a target and a generalization capability is increased, and under limited sample information, an optimal solution of the problem can be acquired.

The invention discloses an image segmentation method based on minimum spanning trees and a statistical learning theory. On the basis of image map model construction, a bottom-up combination strategy is adopted, and the process of generating a plurality of minimum spanning trees is finished according to statistical learning theory-based synthesis criterions designed by the invention by using a minimum spanning tree algorithm. The segmentation method can effectively utilize regional statistical characteristics at the same time of meeting global optimum, is suitable for the high-resolution segmentation of various images, and achieves relatively higher noise immunity and relatively better segmentation effects on texture regions; and in the method, good regional boundaries can be obtained simultaneously.

The invention discloses a structural risk minimization based weighted least squares power system state estimation method. According to the characteristic that the measurement number of power system state estimation is limited, a structural risk minimization based weighted least squares estimation model is provided through a statistical learning theory, so that the norm of a residue can be minimized, and meanwhile the confidence interval of a state variable can be minimized. Besides, a detailed solving process of the method is provided. The method accords with the structural risk minimization thought in the statistical learning theory, can obtain an estimation result more approximate to the true value of the state variable under limited measurement conditions, and has good engineering application prospects.

The invention discloses a structure damage identification method based on independent components and a support vector machine. The method includes: firstly, adopting an ICA to extract an independent component and a hybrid matrix which represent structural characteristics; calculating the front four-order statistical characteristics of each independent component, deforming the hybrid matrix into asingle column vector, taking the combination of the front four-order statistical characteristics and the column vector after the hybrid matrix is deformed as a structural damage characteristic index,and finally inputting the structural damage characteristic index into a support vector machine (SVM) classifier based on a statistical learning theory to carry out damage identification. The method provided by the invention has more effective identifiability and repeatability.

The invention discloses a compressed sampling sensing matrix disturbance optimization model construction method based on statistical learning. The method comprises the following steps: setting a measurement matrix mathematical model phi with disturbance and a sparse matrix mathematical model psi with disturbance; obtaining a disturbance mathematical model A of the sensing matrix according to phi and psi, wherein the formula is A=[phi]*[psi]; obtaining a compressed sensing sparse signal model according to A, n being additive white Gaussian noise, theta being a sparse vector, and y being an observation data vector; and establishing a robust compressed sensing optimization function according to the sparse vector theta, and enabling f(theta) to be equal to ||theta||<1>, converting f(theta) into G (y0, y), and obtaining a sensing matrix disturbance optimization model. The statistical learning theory is applied to compressed sensing, and a robust compressed sensing reconstruction optimization mathematical model for resisting disturbance uncertainty in measurement is established, so that the influence of external environment and internal interference on the measurement uncertainty in thesampling process is reduced, and the accuracy of signal acquisition is improved.

The invention discloses a leaf wetting time monitoring method and system, and the method and system can calculate the wetting time of a leaf more accurately. Moreover, the method and system does not need to carry out the adjustment of equipment in the whole process according to the growth condition of plants, and are simper in calculation process. The method comprises the steps: S1, collecting a fluorescence image of the leaf; S2, carrying out the clustering segmentation of the fluorescence image through employing K-means clustering, carrying out the binarization of a clustering segmentation result, and obtaining a binary image; S3, carrying out the correction of the binary image in a mode of on / off alternate filtering; S4, enabling the preset water drop shape features and size to serve as the judgment standard for judging whether the leaf is wetted or not, employing a support vector machine based on the statistical learning theory to serve as a classifier for distinguishing whether the fluorescence image of the leaf is wetted or not, carrying out the recognition of the corrected image, and obtaining the number of images of the wetted leaves; S5, calculating the wetting time of the leaf according to the number of images of the wetted leaves and the photographing time interval.

The invention discloses a machine learning method for reducing the inconsistency between traditional Chinese medicine subjective questionnaires. The machine learning method comprises the following steps that 1), subjective questionnaire data is vectorized, wherein a subjective questionnaire consists of questions, weight and point values, and the vectorization allows a structure of the questionnaire to be converted into a vector; 2), a consistency target of questionnaire groups is defined and expressed; a contradiction function C(x) is defined to express the consistency between the questionnaire groups; the contradiction function takes a value obtained by converting a point value of the questionnaire group as an input, wherein negative correlation serves as the contradiction function in the process, and accords with a statistical learning theory; 3), main subjective questionnaires such as an NPQ (Nonverbal Personality Questionnaire), an MPQ (Mcgill Pain Questionnaire) and an SF-36 (Short Form-36) used by traditional Chinese medicine are subjected to consistency analysis, wherein each of the NPQ and the MPQ has a sub-questionnaire, and the SF-36 has eight sub-questionnaires; the following objective function of the consistency is defined according to the contradiction function in Step 2); and 4), the objective function is solved. The machine learning method reduces the inconsistency between results of different traditional Chinese medicine treatment effect evaluation questionnaires, and improves the accuracy of the evaluation on a traditional Chinese medicine treatment effect.

The invention discloses an automatic medical image processing system. Through close combination of image data mining theoretical research and practical application, the mainstream technology of current medical image data mining is discussed. The method is based on improvement of a statistical learning theory, a decision tree classification algorithm, an association rule algorithm and a rough set decision tree algorithm. According to the association rule classification method, the problems existing in association rule classification are solved, formats of multi-source medical equipment data arepackaged, a unified processing interface is adopted to form a normalized data directory, redundant features are removed, attribute reduction is carried out, the classification efficiency is improved,and the processing time is shortened. According to the construction of the medical knowledge graph, a key technology can be applied to computer-aided diagnosis, and a medical image automatic processing system based on an image data mining technology is constructed.

The invention belongs to the technical field of compressors, and particularly relates to a fault diagnosis method of an air valve for a reciprocating compressor based on a statistical learning theory. The method comprises the following steps: determining the degree of each data point belonging to a certain cluster; optimizing the objective function through iteration, obtaining the degree of membership of each clustering center and each data point to each class, and classifying samples; realizing the data classification by searching an optimal classification hyperplane; obtaining a dual problem and a classification decision function; constructing a kernel function classifier by utilizing a radial basis kernel function, and solving a dual problem by adopting a sequence minimum optimization algorithm; solving a clustering center of each subset; obtaining M-1 support vector machines SVM1,..., SVMM-1, and forming a root node and an intermediate node of the binary tree; training the support vector machines one by one, optimizing a radial basis kernel function parameter gamma and a penalty parameter C, and realizing fault diagnosis of the air valve for the reciprocating compressor. Therefore, accurate diagnosis of the fault of the air valve can be efficiently realized.

The invention discloses a medical automatic image processing system. The image data mining theoretical research is tightly combined with the practical application, and the mainstream technology of current medical image data mining is discussed; the method is based on improvement of a statistical learning theory, a decision tree classification algorithm, an association rule algorithm and a rough set decision tree algorithm. The invention further discloses an association rule classification method, so problems existing in association rule classification are solved; the format of the multi-sourcemedical equipment data is packaged, and a unified processing interface is adopted to form a normalized data directory, so the classifying efficiency is improved and the processing time is reduced; meanwhile, the key technology can be applied to computer-aided diagnosis through the construction of the medical knowledge graph, and the medical automatic image processing system based on the image data mining technology is constructed.

The invention discloses a method and system for monitoring the wetting time of leaves, which can calculate the wetting time of leaves more accurately, and does not need to adjust equipment according to the growth and development of plants in the whole process, and the calculation process is relatively simple. The method includes: S1, collecting fluorescence images of leaves; S2, clustering and segmenting the fluorescence images using K-means clustering, and binarizing the results of the clustering and segmentation to obtain binarized images; S3, The binarized image is corrected by means of open and close alternate filtering; S4, the preset water droplet shape feature and area size are used as the criterion for judging whether the leaf is wet or not, and the support vector machine based on statistical learning theory is used to distinguish the leaf A classifier for whether the fluorescence image is wet or not, identifies the corrected image, and obtains the number of images of leaves being wet; S5, calculates the wet time of the leaves according to the number of images of the leaves being wet and the time interval of taking pictures.

The invention discloses an opportunity spectrum access method under the condition of limited energy. The method comprises the following steps: establishing a perception access energy consumption channel model; checking residual energy of the cognitive user, and determining a channel set to be perceived; sensing the channels to obtain a sensing idle channel set; enabling the secondary user to access all channels which are perceived to be idle; and enabling the secondary user to update the channel statistical information and the energy information. The perceptual access energy consumption channel model established by the method is closer to reality, the problem is abstracted into an energy-limited multi-arm bandit model based on sequential decision and statistical learning theories, and a channel perceptual access strategy is provided by utilizing a knapsack problem in the optimization problem. Besides, the algorithm designed by the invention has a nonlinear relationship with the magnitude of the user energy in simulation and theory, which indicates that the larger the user energy is, the longer the sensing time is, the more the user can sense the access to a better channel, and better benefits can be obtained.

The invention relates to a method for analyzing the service state of numerical control equipment, belonging to the technology for monitoring the service state of major equipment and forecasting the service life of the major equipment. The method comprises the following steps: firstly, recognizing the service states of a plurality of the major parts of the numerical control equipment through the characteristic physical quantity acquired by a multisensor; then, forecasting the service state of the complete machine through a classification model of a support vector machine established by the statistical learning theory; and calculating the residual service lives of the major parts and the complete machine through a 'hidden semi-Markov' random process model. The method of the invention not only can be used for recognizing the current running states of the parts but also can be used for forecasting the residual service life of the parts. The current running state and the residual service life of the complete machine are obtained by the classification forecasting method of the support vector machine according to the operation result of each part. The invention provides a new method for decision support of preventive maintenance.

The invention discloses a spatial information change detection and classification method based on spatial information and statistical learning. By comprehensively collecting spatial information, analyzing and determining the changing characteristics and classification rules of various spatial elements in a vector layer, combined with statistical learning theory, Use the support vector machine model for sample training, and then detect and classify spatial information changes; through the above method, the present invention can realize comprehensive and rapid detection and classification of spatial information changes, improve detection accuracy, and the method is simple and applicable to When the number of samples is small, the scope of application is wider.

The present invention relates to the wind speed prediction method and device based on G-L mixed noise characteristic nuclear ridge regression technology, the method comprises the following steps: 1) obtain wind speed data set D1, utilize Bayesian principle, obtain the loss function of Gauss-Laplace mixed noise characteristic; 2) Utilize statistical learning theory and optimization theory, combined with the loss function in step 1), establish the original problem of the kernel ridge regression model based on Gauss-Laplace mixed noise characteristics, derive and solve the dual problem of the kernel ridge regression model; 3 ) Determine the optimal parameters of the dual problem of the kernel ridge regression model, select the kernel function, and construct the decision function of the kernel ridge regression model; 4) Construct the wind speed forecast model of the kernel ridge regression model, and use the forecast model to forecast and analyze the wind speed value. The device includes a loss function acquisition module, a dual problem solving module, a decision function construction module and a wind speed forecast module. The invention can meet the requirements of wind speed forecast accuracy in practical applications, such as wind power generation, agricultural production and the like.

The present invention relates to the characteristic of mixed noise based on G-L v ‑Support vector regression machine wind speed forecasting method, the method includes the following steps: 1) Acquire wind speed data set D l , using the Bayesian principle to obtain the empirical risk loss function of Gauss-Laplace mixed noise characteristics; 2) Using statistical learning theory and convex optimization technology, combined with the loss function in step 1), to establish a Gauss-Laplace mixed noise characteristic based on v ‑The original problem of the support vector regression model, using the Lagrange multiplier method to derive and solve the v ‑Support Vector Regression Model Dual Problem; 3) Determine the v ‑Support vector regression model optimal parameters for the dual problem, select the kernel function, and construct the v ‑Decision function of support vector regression model; 4) Construct the v ‑Support vector regression model for wind speed forecasting model, using this forecasting model to predict and analyze wind speed values. The method includes an empirical risk loss function acquisition module, a dual problem solving module, a decision function construction module and a wind speed forecasting module. The invention can meet the requirements of wind speed forecast accuracy in practical applications, such as wind power generation, agricultural production and the like.

The invention discloses a method for measuring and calculating water requirement of regional crops, comprising the following steps of: selecting the measuring and calculating area, invoking DEM, crop planting structure, soil type, weather factors and water requirement data of the crops about the measuring and calculating from a system database; interpolating the terrain factor data and the weather factor data to obtain a grid database set of regional impact factors; extracting comprehensive factor date having an impact on the crop water requirement from each terrain factor data, soil type data and weather factor data; establishing nonlinear mapping relationship between the comprehensive factor date and the crop water requirement to acquire the water requirement data of the different crops; the invention can use diverse space processing methods aiming at diverse impact factors, sufficiently takes account of the space variability of each impact factor, and through the change process of original data and the application of statistic theories, the invention can efficiently reduce the operation workload, improve the data analysis efficiency, and enhance the measuring and calculating precision for the water requirement of regional crops on condition of complicated terrain and sparse site.

The invention discloses a structural risk minimization based weighted least squares power system state estimation method. According to the characteristic that the measurement number of power system state estimation is limited, a structural risk minimization based weighted least squares estimation model is provided through a statistical learning theory, so that the norm of a residue can be minimized, and meanwhile the confidence interval of a state variable can be minimized. Besides, a detailed solving process of the method is provided. The method accords with the structural risk minimization thought in the statistical learning theory, can obtain an estimation result more approximate to the true value of the state variable under limited measurement conditions, and has good engineering application prospects.