50 results about "Nonlinear support vector machine" patented technology

The invention discloses a quick feedback analyzing system in a tunnel constructing process. The system adopts a scheme: understanding currently adopted designing construction parameters; establishing a tunnel excavation three-dimensional finite element numerical grid calculation model; acquiring surrounding rock layering and convergent displacement monitoring information after a tunnel is excavated; establishing a non-linear support vector machine model; fixing an anchoring parameter according to the actual construction parameter, and optimally identifying rock mechanic parameters by adoptinga differential optimization algorithm; optimizing the construction parameter of an anchoring scheme by adopting a differential evolution algorithm; and optimizing the rock mechanic parameters by calling the differential evolution and optimization algorithms to further solve the construction parameter of the anchoring scheme, and outputting the construction parameter of the optimized anchoring scheme as a construction scheme through a computer display screen to guide the constructors to construct. The quick feedback analyzing system ensures that the monitoring information is used for optimizing the anchoring parameter while being used for identifying the surrounding rock parameters, so that the dynamic information construction is improved to a level of quantitative analysis.

The invention provides an image object recognition method based on SURF (Speed Up Robust Feature), comprising the following steps: first, preprocessing images; second, extracting SURF corners and SURF descriptors of the images to describe the features of the images; third, processing the features through PCA data whitening and dimension reduction; establishing a bag-of-visual-words model through Kmeans clustering based on the features after processing, and using the bag-of-visual-words model to construct a visual vocabulary histogram of the images; and finally, carrying out training by a nonlinear support vector machine (SVM) classification method, and classifying the images to different categories. After classification model building of different images is completed in the training phase, the images tested in a concentrated way are detected in the testing phase, and therefore, different image objects can be recognized. The method has excellent performance in the aspects of recognition rate and speed, and can reflect the content of images more objectively and accurately. In addition, the classification result of an SVM classifier is optimized, and the error rate of judgment of the classifier and the limitation of the categories of training samples are reduced.

The invention discloses a medical insurance fraud detection method based on multiple features. The method comprises: aimed at treatment histories of all patients suffered from the same disease in medical insurance declaration data, in combination with medicine classification knowledge, adopting probability statistics, mixture Gaussian modeling, feature fusion and other techniques to extract multiple secondary feature data with higher distinction degree; then, vectorizing the treatment histories of the patients based on the secondary feature data; and then, after carrying out clustering analysis on treatment data marked with 'normal', adopting a non-linear support vector machine classification technique to establish multiple classification hyperplanes for each type of normal treatment data subjected to clustering and treatment data marked with 'fraud', so that the fraud detection can be carried out on non-marked medical insurance data. The method can be used for quickly and effectively detecting the fraud data existent in the medical insurance data, and has relatively high accuracy.

The invention discloses a network access type decision method. The method comprises the following steps: acquiring a data transmission rate when a visible light communication channel is unshielded under the current state of a user side; according to a known channel blocking parameter of the previous state and the data transmission rate when the visible light communication channel at the current state is unshielded, determining the access type with the maximum equivalent data rate as the access type of the user terminal at the current state according to a pre-trained nonlinear support vector machine model. The invention further provides a network access type decision device, a network switching control device and a computer readable storage medium. In a hybrid indoor wireless communicationnetwork environment composed of visible light communication and the traditional radio frequency communication, the user can comparatively accurately select the access type with the equivalent data rate at the current state under the condition that the actual channel block parameter is unknown, thereby effectively responding to the adverse influence caused by frequent switching and shielding blocking, and the demands on the transmission rate and the communication quality by the user are satisfied.

Non-intrusive assessment of fatigue in drivers using eye tracking. In a simulated driving experiment, vigilance was assessed by power spectral analysis of multichannel electroencephalogram (EEG) signals, recorded simultaneously, and binary labels of alert and drowsy (baseline) were generated for each epoch of the eye tracking data. A classifier and a non-linear support vector machine were employed for vigilance assessment. Evaluation results revealed a high accuracy of 88% for the RF classifier, which significantly outperformed the SVM with 81% accuracy (p<0.001). In a simulated driving experiment, the simultaneously recorded multichannel electroencephalogram (EEG) signals were used as the baseline. A random forest (RF) and a non-linear support vector machine (SVM) were employed for binary classification of the state of vigilance. Different lengths of eye tracking epoch were selected for feature extraction, and the performance of each classifier was investigated for every epoch length. Results revealed a high accuracy for the RF classifier in the range of 88.37%-91.18% across all epoch lengths, outperforming the SVM with 77.12%-82.62% accuracy. A feature analysis approach was presented and top eye tracking features for drowsiness detection were identified. A high correspondence was identified between the extracted eye tracking features and EEG as a physiological measure of vigilance and verified the potential of these features along with a proper classification technique, such as the RF, for non-intrusive long-term assessment of drowsiness in drivers.

The invention discloses an accelerometer temperature compensation method, and relates to the technical field of sensor soft compensation. The method comprises the following steps that in different acceleration environments, acceleration values sensed by an accelerometer in the different temperature conditions in each acceleration environment are selected as sample data; a particle swarm algorithmis used to optimize a parameter of a support vector machine to obtain an optimal parameter, and the optimal parameter is to establish an optimal nonlinear support vector machine; and the optimal nonlinear support vector machine is used to predict the sample data, a nonlinear compensation model is established via the support vector machine of a PSO-SVM particle swarm regression machine, and the test accelerations of the accelerometer at different temperatures are compensated nonlinearly. The calculating accuracy is improved, the instantaneity is high, and the speed measuring application range of the accelerometers is widened greatly.

The invention relates to a pavement marking automatic extraction method and device, and the method comprises the steps: firstly selecting various types of pavement point cloud samples, generating a corresponding feature vector according to the three-dimensional information of pavement point cloud, carrying out the training of a nonlinear support vector machine through the generated feature vector,and obtaining a classification model; and then processing the to-be-classified road surface point cloud, generating a corresponding feature vector according to the three-dimensional information of the road surface point cloud, and classifying the to-be-classified point cloud by utilizing the classification model. According to the method, the three-dimensional information of the point cloud coordinates is utilized, the feature vectors are calculated according to the three-dimensional information, and the calculated feature vectors are classified by combining a nonlinear support vector machinemethod, so that the error extraction rate of the road surface non-marked line point cloud can be effectively reduced, and the calculation amount is greatly reduced.

The invention discloses a cordyceps sinensis detection method based on self-encoding feature learning. The cordyceps sinensis detection method comprises the following steps of (1) collecting a series of images containing cordyceps sinensis; (2) extracting cordyceps sinensis and other backgrounds from the image, and making positive and negative samples with same size; (3) training a self-encoding model by the extracted sample images, and obtaining self-encoding model parameters; (4) encoding the samples through the self-encoding model; (5) classifying and training the obtained encodes and sample types by a nonlinear support vector machine, and obtaining classifying model parameters; (6) collecting the to-be-detected cordyceps sinensis images, and blocking under multiple scales; (7) encoding each image by the self-encoding model, and using the nonlinear support vector machine to classify and record positions; (8) removing the detected coinciding areas, and identifying all non-coinciding areas onto the to-be-detected images. The method can be used for automatically detecting the cordyceps sinensis in the environment under the complicated background.

The invention relates to the technical field of sensor software compensation and specifically discloses a sensor nonlinear compensation method. The method comprises the steps of selecting resistance values corresponding to a plurality of pieces of temperature sensed by platinum resistor in a certain temperature interval as sample data; optimizing parameters of a support vector machine through adoption of a bat algorithm, thereby obtaining the optimum parameters, and establishing the optimum nonlinear support vector machine through adoption of the optimum parameters; and predicting the sample data through utilization of the optimum nonlinear support vector machine. According to the method, a nonlinear compensation model is established through application of a bat algorithm (BA-SVM) supportvector machine, so the platinum resistor nonlinear compensation method is realized. Nonlinear characteristics of a platinum resistor sensor is compensated. A compensated system can be processed according to linear characteristics. The computing accuracy is improved. The timeliness is good. An platinum thermistor temperature measurement application range is greatly expanded.

The invention relates to a nuclear power system fault diagnosis method and system. According to the nuclear power system fault diagnosis method and system, through a trained nonlinear support vector machine model, the fault category of each subsystem in a nuclear power system and the fault occurrence probability corresponding to the fault category can be accurately obtained; after a normalizationcoefficient is determined according to the fault occurrence probability in a constructed two-dimensional fault probability matrix, a fault probability value corresponding to each fault is determined according to the normalization coefficient; and the difference value between a first large fault probability value and a second large fault probability value in the probability values after descendingsort is compared with a set threshold value to obtain the fault diagnosis result of the nuclear power system, so that the adaptability of the whole nuclear power system fault diagnosis method and system is improved while the diagnosis accuracy is improved.

The invention provides an application program management and control method and device, a medium and an electronic device. Historical characteristic information xi is obtained by detecting that an application program enters a background, a nonlinear support vector machine algorithm is adopted to generate a training model, accordingly the current characteristic information s of the application program is introduced into the training model, thus whether the application program needed to be closed or not is judged, and the application program is intelligently closed.

The invention discloses human body behavior recognition based on a logarithmic Euclidean space BOW (bag of words) model, and belongs to the technical field of digital image processing. The recognition comprises the steps: firstly enabling an input video to be divided into video segments which have a fixed length and are overlapped; secondly cutting each video segment into space-time cubic blocks which have the fixed size and are partly overlapped; thirdly extracting a gradient and a light stream feature covariance or a shape feature covariance of each space-time cubic block, and carrying out the dimension reduction of a covariance matrix through employing a symmetric positive definite matrix dimension reduction method. The method carries out the logarithmic change of the covariance matrix, extracts the triangular features of a logarithmic covariance matrix, and converts the triangular features into a logarithmic Euclidean space vector. The method carries out the behavior modeling for the logarithmic Euclidean space through employing the BOW model, carries out the clustering of behavior characteristics through employing spectrum clustering to generate a codebook, and codes the behavior characteristics through employing the LLC (Locality-constrained Linear Coding) technology. A nonlinear support vector machine is used for the training, recognition and classification of the behavior characteristics. The method is used for the recognition of human body behaviors, and is great in robustness.

The invention discloses a multi-variable monitoring method and system based on a Savitzky-Golay filter. The method comprises: obtaining the historical operation data of a multi-variable system, and marking the normal or abnormal condition of the historical operation data, wherein the historical operation data contains a plurality of process variables indicating a system operation state; using theSavitzky-Golay filter to filter the historical time sequence data of each process variable to obtain a filtered signal amplitude and each order derivative; making the filtered signal amplitude and each order derivative of each process variable form a characteristic space; and training and acquiring a nonlinear support vector machine model based on historical data in the characteristic space so asto determine whether to give an alarm, and realizing multi-variable monitoring. The method has high accuracy, the disadvantage of a traditional univariate alarm threshold design method is overcome, the finiteness of the monitoring system is effectively improved, and the number of false and missing alarms is greatly reduced.

The invention relates to the field of circuit breaker diagnosis, and particularly discloses a contact ablation fault estimation method of a high-voltage circuit breaker. The contact ablation fault estimation method comprises the steps of acquiring a range-time curve of the circuit breaker; obtaining a resistance-range curve according to a resistance-time curve and the range-time curve; acquiring acontact ablation state parameter numerical value of the circuit breaker according to each resistance-range curve; optimizing a parameter of a support vector machine by particle swarm optimization toobtain an optimal parameter, and building an optimal non-linear support vector machine by employing the optimal parameter; building sample data; training the optimal non-linear support vector machineby employing the sample data, inputting the resistance-range curve, and outputting the corresponding contact ablation state parameter numerical value to obtain the non-linear support vector machine capable of estimation; and forecasting the resistance-range curve of the circuit breaker to be estimated by employing the trained non-linear support vector machine. By the method, the contact ablation fault of the high-voltage circuit breaker can be accurately estimated.

The invention discloses a fatigue monitoring system and method fusing myoelectricity and electrocardiosignals. The system comprises a surface electromyogram signal electrode module, a surface electromyogram signal acquisition and conversion module, a surface electromyogram signal data processing module, an electrocardiosignal acquisition and conversion module, an electrocardiosignal data processing module and a nonlinear support vector machine algorithm data fusion module. The method comprises the steps that firstly, driver surface electromyogram signals are collected, then signal amplification, filtering and A/D conversion are carried out, and then analysis is carried out to obtain characteristic parameters of the driver surface electromyogram signals; then, electrocardiosignals of a driver are collected, amplification, filtering and A/D conversion are carried out, and then characteristic parameters of the electrocardiosignals of the driver are obtained through analysis; and feature layer fusion is performed on the surface electromyographic signal feature parameters and the electrocardiographic signal feature parameters of the driver to obtain a fatigue feature vector, and the fatigue condition of the driver is judged by judging whether the feature vector conforms to fatigue features or not. The risk of fatigue driving of the driver is reduced.

The invention provides an application program control method and device, a medium and electronic equipment. The method comprises the steps of obtaining historical feature information xi, adopting a BPneural network algorithm for generating a first training model, adopting a nonlinear support vector machine algorithm for generating a second training model, when it is detected that an application program enters a background, substituting current feature information of the application program into the first training model to obtain a first closing probability value, when the first closing probability value is within a hesitating region, inputting the current feature information s of the application program into the second training model for calculation to obtain a second closing probabilityvalue, judging whether the application program needs to be closed or not, and intelligently closing the application program.

The invention relates to a rapid nondestructive testing method for a large-size composite material, which is characterized by comprising the following steps: randomly sampling according to a standard component to obtain a rapid sparse representation vector of a defect-free reference signal sample set; performing fast sparse characterization, the sparse characterization of the reference signal and the sampling signal adopting parallel noise reduction and sparse representation of signal singularity measurement based on a wavelet domain modulus maximum value, feature clustering of a nonlinear support vector machine, an adaptive observation model of correlation entropy and normalization of a Sigmoid function; using a maskless compressed sensing (CS) topological distribution optimization strategy of a correlation entropy adaptive model, and characterizing a result and the correlation entropy adaptive model rapidly; judging the homogeneity and heterogeneity of signals through weighting, designing a structured measurement matrix easy to store under a CS theoretical framework in combination with a block random form of a structured thought, and providing effective data for reconstruction after rapid detection. The method is closely related to detection object parameters, the implementation method is simple and easy to implement, and the engineering practicability is high.

The invention relates to a vehicle-mounted fatigue detection method based on a multi-scale binary mode, which solves the technical problem of low detection precision. The method comprises the following steps: dividing a training sample image into a plurality of non-repeated sub-regions, and carrying out feature extraction by using the multi-scale local binary mode to obtain multi-scale local binary image features; performing discrete Fourier transform on the multi-scale local binary image features to obtain a histogram Fourier feature vector of a multi-scale binary mode; step 4, connecting histogram Fourier feature vectors which form a multi-scale binary mode, for representing image features, selecting a kernel function, carrying out classification training on MLBP features of sample images through a nonlinear support vector machine, and obtaining trained SVM classification models and parameters. The problem is well solved, and the method can be used for vehicle-mounted fatigue detection.

The invention discloses exhaled gas detection equipment, which comprises a gas collection device used for collecting exhaled gas; a mass spectrometer used for analyzing the exhaled gas to obtain a VOCs mass spectrum; a processor used for extracting VOCs content data from the VOCs mass spectrum, inputting the VOCs content data into a pre-prepared classification model, and outputting to obtain a detection result, wherein the classification model is a nonlinear support vector machine model, and a penalty factor c and a gamma parameter of the classification model are obtained through genetic algorithm optimization; and a display used for displaying the classification model and the detection result. The classification model for detecting the exhaled gas is obtained based on the nonlinear support vector machine, and the penalty factor c and the gamma parameter are obtained through genetic algorithm optimization, so that the stability and the accuracy of the classification model are greatly improved, and the equipment can be applied clinically.

The invention provides a video quality evaluation method and system combining a support vector machine and fuzzy reasoning, and the method comprises the steps: firstly extracting main influence factors of video quality, classifying the video quality through an optimized nonlinear support vector machine model, and then dividing the classified influence factors into two groups, and carrying out fuzzy reasoning according to the rule corresponding to each classification, and weighting reasoning result value to obtain a final objective value. Experimental results show that the method can effectively improve subjective and objective similarity of video quality evaluation.