147 results about "Svm regression" patented technology

The invention discloses a genetic algorithm-based support vector regression shipping traffic flow prediction method, comprising the following steps: (1) reducing the dimensional number of factors possibly generating influences on shipping traffic flow by a weighted principal component analysis method, and selecting influencing factors with higher cumulative contribution rate; (2) carrying out normalization preprocessing on original vessel traffic flow time series data to generate a data set and then grouping; (3) selecting a kernel function to determine support vector machine (SVM) regression parameters; (4) constructing a support vector regression prediction model optimized by a genetic algorithm; (5) inputting the data set to generate a prediction function; and (6) predicting according to the prediction function generated in step (5), evaluating and analyzing prediction error, and if the error is relatively large, returning to step (2) and regulating the parameters again, predicting once again. The method of the invention has the advantages of higher prediction accuracy and higher stability of prediction accuracy.

The invention discloses a power equipment current-carrying fault trend prediction method based ona least squares support vector machine. The method provided by the invention comprises the steps of employing historical temperature data to train an LS-SVM regression model, and employing a PSO optimization algorithm to adjust two parameters of the model, namely nucleus width sigma and punishment parameter gamma; employing a PCA algorithm and a K-means clustering algorithm to real-time analyze the temperature of equipment contacts to find contacts with abnormal temperature rising, and using the temperature value asan initial value sequence of prediction;and finally employing the regression model obtained by training to predict the temperature value of the initial value for a long term and for a short term, and analyzing the highest point the contact temperature may reach and the time when the contact temperature reaches the highest point. Through predictive analysis based on PSO-LSSVM, fault development trend of equipment contacts is actively controlled, so the time for timely measures and ensuring the safe operation of power grid is bought. The method provided by the invention can be widely used in the field of power equipment forecast alarm protection.

The invention discloses an indoor passive positioning method based on channel state information and a support vector machine. The method comprises the following steps: firstly preprocessing the acquired channel state information data, performing de-noising and smoothness through the adoption of a density-based spatial clustering of applications with noise and a weight-based moving average algorithm, and then using the principal component analysis algorithm to extract the features. The data after the preprocessing and feature-extracting can reflect the signal change more accurately and the dimension is greatly reduced. The passive positioning adopts two-stage positioning. In the training stage, the large positioning space is divided into sub-regions, the support vector machine classification and regression model is established for each sub-region so as to acquire a statistic model for accurately representing the nonlinear relationship between the position and the signal. The two-stage positioning firstly determines the sub-regions through the classification of the support vector machine, and the precision position is determined in the sub-region through the regression of the support vector machine. The method disclosed by the invention has the beneficial effects that the passive positioning can be performed in the absence of the active participation of the target, and the indoor positioning precision is improved to sub-meter level.

The present invention claims to protect a shared bicycle management and decision-making method. The method comprises: determining a delivery area, collecting parameter data, using a fuzzy comprehensive evaluation model of DEA to construct a membership matrix to determine an indicator level of the delivery area, and supplemented by multi-objective planning, determining the delivery number of bicycles in the delivery area and the delivery number of bicycles in different delivery areas; and using SVM regression and using the collected parameters to determine the function relationship of the delivery number of shared bicycles over time through the mapping of the Gaussian kernel function in the high-dimensional space, and to determine the delivery number of bicycles at different times in the same area. According to the method provided by the present invention, a quantitative decision-making management method is proposed, a FICO based personal credit evaluation system is established, bicycleuse by users is standardized, and through multi-objective planning, an optimal dispatching and repair center is set according to the different delivery points in the region, so that the operation andmaintenance costs can be reduced.

The invention belongs to the technical field of equipment state early warning, and particularly relates to an evaluation value calibration method of an equipment intelligent early warning system based on an SVM regression model. The evaluation value calibration method of the equipment intelligent early warning system can remove deviation of an abnormal detecting point with respect to the evaluation values of other normal measuring points in real time sequentially through establishment of an equipment intelligent early warning model, influence relation curve fitting, identification of abnormal measuring points and calibration of interfered evaluation values, has good universality and can be grafted to other regression algorithms for performance optimization. When the evaluation value of an interfered measuring point of equipment is adjusted on line, no excessive delay effect is generated, and a real-time early warning capability of the model for the equipment is ensured; discrimination and analysis for abnormal measuring points can be carried out in real time, so that the real-time performance of judgment is ensured, and online state analysis for equipment data can be carried out; and evaluation values without interference of the normal measuring points can be output steadily. According to the evaluation value calibration method of the equipment intelligent early warning system, the reliability of the early warning model is higher, and the life cycle is long.

The invention discloses a multi-model integrated flood forecasting method based on propagation time clustering analysis and belongs to the technical field of hydrologic forecasting. The method includes adopting a derived dynamic time warp matching method for floor process similarity analysis; estimating flow propagation time of each station in the upstream and the downstream; decomposing a sample into a plurality of clusters by performing clustering analysis on the flow propagation time; respectively building an SVM regression model for each sub-flow sequence to simulate the flood forming process; combining the sub-models into a comprehensive model. Comparing a comprehensive forecasting result acquired by the method with a model forecasting result acquired by a single model under conventional conditions and on the basis of flow clustering, the comparison result shows that the comprehensive model is better in comprehensive performance.

The invention discloses a forecasting method for a solar photovoltaic electricity generation amount based on the SVM (support vector machine) - Markov combination method. The method comprises the following steps of (1) selecting the solar radiant strength, daily maximum temperature, relative humidity and a wind speed as warning factors; (2) collecting a certain quantity of sample data according to the warning factors; (3) primarily establishing an SVM regression forecast model, carrying out training by utilizing the sample data, and determining an SVM model structure; (4) carrying out primary forecast of the photovoltaic electricity generation amount according to the SVM model structure obtained from the step (3); (5) carrying out rectification on a forecast result by applying the Markov method; (6) obtaining the forecast result. According to the forecasting method, the SVM is adopted to carry out the regression forecasting analysis, rectification on the forecast result is carried out through the Markov method, the method is coincided with the characteristics of photovoltaic electricity generation, advantages of the forecasting method and the Markov method are complemented, therefore a more accurate forecasting result is obtained, and reliable forecast on the photovoltaic electricity generation amount is realized.

The invention relates to a GPS/INS integrated navigation method based on unscented Kalman filtering. The method is mainly characterized in that the method comprises the following steps that INS data and corresponding GPS data are combined together, and an unscented Kalman filtering method is adopted to build a GPS/INS integrated navigation model to obtain the online optimal value of a GPS; the INS data are used as training data and are combined with the online optimal value of the GPS as a training target to build an SVM regression model, and the SVM regression model is used for predicting the off-line navigation data of the GPS. According to the method, the precision of the GPS is usually combined with the reliability of an inertial navigation system to achieve the information fusion of the filtering, the GPS and an INS, and when the GPS and the INS are used together, a precise filtering result can be obtained; an SVM is adopted to train INS data obtained when GPS signals are lost, the parameters of a kernel function and a penalty function in an SVM algorithm are optimized through a simulated annealing algorithm, and the precision of the off-line GPS is kept close to that of the online GPS.

The invention relates to a grain condition forecasting and early warning method and system based on the SVM. The method includes the steps of setting multiple parameters which affect security levels of the grain condition, forming a standardized historical data sample, setting up a forecasting model based on an SVM regression model, collecting data of all the parameters, obtaining a forecasting result of the security level of the grain condition through the forecasting model, judging whether the change trend of the forecasting result of the security level of the grain condition is normal or not, if yes, sending the forecasting result of the security level of the grain condition to an upper computer, and if not, sending out an alarm signal. According to the method and the system, comprehensive analysis can be conducted on the collected grain condition according to the set forecasting model, and therefore the change trend of the security level of the grain condition can be forecasted; when the grain condition is abnormal, an alarm is given so that administrative staff can be prompted to do preparation work for improving the grain condition in advance, the timely foundation is provided for the control strategy of a grain condition monitoring and control system, and reliability of the monitoring and control system is improved.

The invention discloses a power material demand prediction method based on text information extraction. The power material demand prediction method includes a two-step algorithm of power material demand prediction, wherein the first step is used for processing a preliminary design document based on the text information extraction technology, and extracting the engineering attribute information which has important value for predicting the demand quantity of main equipment to realize the structural expression of the preliminary design document, and then realizing the requirement prediction of the main equipment by utilizing an SVM regression algorithm. In the second step, the dense vector expression of a primary design document is learned through a convolutional neural network by utilizing atext classification technology, the demand information of the main equipment is fused with the demand information of the main equipment, and the demand of non-main equipment is predicted through a multi-layer neural network. Compared with the existing calculation, the method can be used for predicting various types of materials. The prediction data tend to be actual, the attributes have more expression, and the method has good practicability. The material demand prediction method conforms to actual application requirements, and can be used for predicting the material requirements after the initial design is completed.

The invention belongs to the field of mining safety and relates to a rock-burst acoustic emission predicting method based on a support vector machine (SVM). The rock-burst acoustic emission predicting method includes using SVM regression prediction as a means, selecting energy release value as judgment indicator for predicting rock burst on the basis of test research analysis, using burst tendency physical and mechanic parameters and acoustic emission signal feature parameters as influence factor of rock burst, and generating SVM input vector by extracting sample data according to these parameters; selecting Gauss radial basis function (RBF) as kernel function of the SVM in setup of a SVM prediction model, and using the K-CV cross validation algorithm to optimize and select the optimal penalty parameter C and kernel function parameter g; applying a fuzzy information granulating data processing method into the regression prediction of the SVM so as to predict change trends and change space of the predicted values of the rock burst.

The invention discloses a multi-information fusion modeling method for the shapes of burden surfaces in the burden distribution process of a blast furnace. The content of the multi-information fusion modeling method comprises: establishing a furnace burden stacking model according to hydromechanics and statistics rules, adopting a mechanism method to deduce a model function for burden surfaces in the burden distribution process of the blast furnace, establishing a parameterized prediction model, solving original burden surface shape parameters, and determining a furnace burden stacking equation; based on the fitting of radar scattered data regressed by a support vector machine, obtaining an original burden surface shape; according to data measured by a radar, solving the original burden surface shape parameters of the prediction model; using the radar to measure the height information of different radii of burden surfaces in multiple point positions of the blast furnace, and adopting the support vector machine which is applied to a regression problem to fit the height scattered points of the burden surfaces so as to obtain a burden surface function curve; determining related parameters according to a burden distribution mechanism relationship and the radar data, and correcting the prediction model according to the parameters; based on simulation results of a discrete element process, correcting the burden surface shape parameters; and obtaining a new burden surface shape function, using the burden surface shape function as an output result, and feeding the output result back as the next original burden surface.

The invention discloses an indoor positioning method based on propagation loss learning. The method comprises steps: a mobile terminal scans all indoor wifi signals in real time, and a logarithmic distance-loss model is built according to the real-time scanning result; historical data are used for learning the distance-loss model, and a machine learning model is built; a training sample set is generated based on the historical data, the built machine learning model is trained, and an SVM regression prediction function for generating a wifi signal transmission attenuation prediction value is obtained; and based on the real-time scanning result of the mobile terminal and in combination with an SVM regression prediction equation and the logarithmic distance-loss model, the accurate position coordinates of the mobile terminal are calculated.

The invention provides a method for counting the numbers of indoor persons on the basis of WiFi (wireless fidelity) channel state information (CSI) and support vector machine (SVM) regression. Specialhardware facilities can be omitted, and the numbers of the indoor persons can be counted only by the aid of existing WiFi wireless networks; CSI data can be denoised by the aid of DBSCAN (density-based spatial clustering of application with noise) algorithms after the CSI data are acquired, then non-zero rates of each subcarrier are obtained by the aid of expansive matrix algorithms and are usedas CSI feature fingerprint samples, accordingly, the influence of great change of signal amplitudes on person number counting can be enhanced, and influence of environmental noise on small change of the signal amplitudes can be reduced; accurate nonlinear dependency relationship models between the numbers of the persons and the CSI feature fingerprint samples can be obtained by the aid of SVM regression algorithms without consideration on complicated indoor environments, and accordingly the purpose of accurately counting the numbers of the indoor persons can be achieved. The method has the advantages that the numbers of the persons can be accurately counted on the basis of the existing WiFi wireless networks, the method is low in cost and high in universality, and the privacy problems canbe solved.

The invention provides a method for performing nuclear magnetic resonance identification on adulterated non-dairy cream in watery cream. The method comprises the following steps: (1) collecting creamsamples and performing plausibility check on the samples; (2) preparing adulterated cream samples by performing plausibility check on the cream samples and collecting 1H-NMR of the samples; (3) establishing a PLS-DA qualitative model and a PCA-SVM regression quantitative analysis model by adopting 1H-NMR data subjected to data processing; (4) collecting 1H-NMR data of the cream samples with unknown adulteration or not, and performing detection by the established qualitative and quantitative models so as to obtain the identification of the unknown cream samples. According to the method, the accuracy of the modeling sample is guaranteed, the condition that the non-dairy cream is adulterated in the watery cream can be detected rapidly, qualitatively and quantitatively, the technology is reliable, the operation is simple and convenient, the adopted model is high in calculation speed and accurate in identification result, and technological support is provided for quality supervision of baking products such as cream cakes.

The invention relates to a near-infrared spectrum rapid evaluation method for rice taste quality, wherein the method is one for food detection by using a near-infrared spectrum. The method comprises the following steps: processing a collected sample; with a near-infrared diffuse reflection spectrum as an original spectrum, preprocessing the spectrum, preliminarily optimizing the characteristic wavelength of the near-infrared spectrum based on a competitive adaptive reweighted sampling algorithm, and performing synchronous optimization on parameters of a support vector machine and the characteristic wavelength of the near-infrared spectrum based on a quantum genetic simulated annealing algorithm to obtain an optimal characteristic wavelength; establishing a regression model by utilizing theoptimal characteristic wavelength, evaluating the precision of the regression model, and rapidly establishing an evaluation model; performing spectral scanning on polished rice needing to be measured, and inputting spectral data into the evaluation model according to the characteristic wavelength to finish quality evaluation. According to the method, the quantum genetic simulated annealing algorithm is applied to synchronously optimize the parameters of the support vector machine and the characteristic wavelength of the near-infrared spectrum, so that the detection precision and efficiency ofthe rice taste quality by the support vector machine regression model are effectively improved.

The invention discloses a method for predicting reservoir productivity by using logging facies combined post-stack seismic attributes, and particularly relates to the field of geophysical exploration.Logging facies modeling standard wells and non-standard wells are divided according to standardized logging curves of all wells in a work area, logging curve principal components of the logging facies modeling standard wells are extracted and subjected to Kmeans clustering analysis, and the optimal clustering number is determined on the basis of an elbow rule to divide logging facies, the same principal components are extracted from the non-standard well logging curve, logging facies are divided by utilizing Kmeans clustering analysis, a logging facies quality model established, reservoir productivity is represented by utilizing logging facies quality, post-stack seismic attributes are extracted, Pearson correlation analysis is performed on the post-stack seismic attributes and the logging facies quality, sensitive seismic attributes are determined, and based on a support vector machine regression algorithm, and a mapping relationship between the sensitive seismic attributes and the logging facies quality is established, and a logging facies quality plane graph is drawn to perform productivity prediction. Accurate prediction of reservoir productivity is realized by using logging facies and seismic attributes, and guidance of oilfield exploration and development is facilitated.

The invention provides methods for constructing and predicting the leaf trait of a woody plant and a photosynthetic characteristic model based on a DNA methylation level, and belongs to the technicalfield of biological analysis. The predicting method comprises selecting important characteristic variable embodying a geographic position difference based on a random forest, screening out 7 leaf characteristic variables, determining an optimal cluster number, and obtaining each group of cluster leaf samples by using an improved FCM clustering algorithm; according to the correlation between variables and the importance of Enzyme digestion combination obtained by a gradient boosted tree, obtaining an important enzyme digestion combination in each group of cluster leaf samples; by using the DNAmethylation level of the enzyme digestion combination as a regression variable, constructing LS-SVM regression prediction model based on Gaussian radial basis function; inputting the DNA methylation level of important enzyme digestion combination to accurately predict a leaf shape factor, leaf area and a net photosynthetic rate. The method is used for predicting the phenotypic characteristic and the photosynthetic characteristic of the woody plant, and screening individuals of woody plants with excellent traits.

The invention provides an SCR denitration system prediction model optimization method based on machine learning. The SCR denitration system prediction model optimization method comprises the followingsteps: S1, collecting the NOx concentration of a boiler outlet in an SCR denitration system and real-time sample data of related indexes influencing the NOx concentration; s2, carrying out dimensionreduction processing by utilizing principal component analysis; s3, establishing a support vector machine model; s4, introducing an exponential decay model to iteratively update the step size value ofthe longicorn beard algorithm, and optimizing vector machine parameters; s5, performing simulation of a support vector machine; and S6, repeating the step S1S5. The invention provides an SCR denitration system prediction model optimization method based on machine learning, which solves the problem that accurate control of ammonia injection quantity is difficult to realize in the existing thermalpower plant; and the invention comprises performing dimension reduction processing on sample data based on a PCA method, iteratively updating a step size value by introducing an exponential decay model, and optimizing by improving a BAS algorithm to obtain optimal support vector machine model parameters, and establishing an optimized support vector machine regression (SVM) model.

The invention belongs to the technical field of power load prediction, and particularly relates to an intelligent prediction method for power generation power of a distributed power supply, which comprises the following steps of: 1, acquiring meteorological sample data including historical daily wind speed and wind direction, and performing normalization processing; 2, dividing the sample data into model training data and inspection data, and performing iterative optimization by using a fruit fly algorithm; step 3, introducing an immune algorithm to optimize parameters of the fruit fly algorithm; 4, using an immune-Drosophila optimization algorithm optimized by immune algorithm to optimize a support vector machine, and establishing an optimized support vector machine prediction model; and5, verifying the effectiveness of the support vector machine prediction model, and obtaining a final prediction result. Through optimization of the immune algorithm, the problem that the fruit fly algorithm is trapped in local optimum is avoided, the global search capability of the algorithm is effectively improved, and the SVM regression prediction precision of the model is improved.

The invention discloses a water quality prediction method based on SVM-KNN. The method comprises the following steps of 1) preprocessing the acquired water quality data; 2) pre-classifying the data byusing a K-nearest neighbor algorithm; 3) establishing a support vector machine regression model, selecting the optimal parameters of a water quality data prediction model through a k-fold cross validation algorithm, and applying the optimal parameters to the model; and 4) training the model, and evaluating the prediction effect by taking the temperature and the pH value as the prediction targets.According to the water quality prediction method combining the KNN and the support vector machine, a water quality time sequence is predicted, the prediction accuracy is improved, the method belongsto the technical field of water quality prediction methods, and the accurate prediction of the water quality is successfully achieved.

The invention discloses a method for predicting net load of a distributed power supply power distribution network. determining impact factors which have impact on predicting photovoltaic power generation, wind power, and load power; inputting the impact factors collected into a data integration model to obtain a reference value of the net load; initially establishing a SVM regression prediction model according to all the impact factors and performing training to obtained samples to obtain a SVM model; performing net load prediction to obtain a predicting result according to the mentioned SVM model; managing the predicting result and scheduling reasonably. By use of the method, the prediction precision is greatly increased; a reasonable predicting result can be output; scheduling is facilitated and cost of power network dispatching is reduced; electric energy production of each power plant is reasonably arranged; operation economical efficiency is increased. The method has high practicality and reliability and is good for safe, stable and reliable operation of power networks.

The invention relates to a correction method for on-line monitoring noisy data of oil chromatography. The method includes the following steps: step 1, collecting data of off-line tests and on-line monitoring of the oil chromatography; step 2, obtaining an optimal combination of significant parameters in a regression model of a support vector machine through a firefly algorithm; step 3, training the support vector machine with the small amount of accurate off-line test data of the oil chromatography obtained, and obtaining the regression model of the support vector machine; step 4, initializing a permissible deviation radius h of the on-line monitoring data, calculating a piecewise function between the off-line tests, and judging whether the on-line monitoring data of the oil chromatography is in a permissible error range of the model; step 5, correcting the on-line data; and step 6, according to the result of correction feedback of the on-site data, adjusting the parameters in the model. When the method is used for correction of the on-line data of the oil chromatography, the effect is stable, the result is accurate, the time is short, and the real-time performance is good, and the method is very suitable for correction of the one-site on-line data of the oil chromatography.