83 results about "Akaike information criterion" patented technology

Ultrasound sound-speed tomography requires accurate picks of time-of-flights (TOFs) of transmitted ultrasound signals, however, manual picking on large datasets is time-consuming. An improved automatic TOF picker is taught based on the Akaike Information Criterion (AIC) and multi-model inference (model averaging), based on the calculated AIC values, to improve the accuracy of TOF picks. The automatic TOF picker of the present invention can accurately pick TOFs in the presence of random noise with average absolute amplitude of up to 80% of the maximum absolute synthetic signal amplitude. The inventive method is applied to clinical ultrasound breast data, and compared with manual picks and amplitude threshold picking. Test results indicate that the inventive TOF picker is much less sensitive to data signal-to-noise ratios (SNRs), and performs more consistently for different datasets in relation to manual picking. The technique provides noticeably improved image reconstruction accuracy.

The invention discloses a high-order multi-stage auto-regressive distributed lag modeling method of thermal error compensation of a numerical control machine. The method is characterized by comprising the following steps of: providing a high-order multi-stage auto-regressive distributed lag modeling formula of a thermal error of a numerical control machine containing a coefficient to be solved, selecting numerical control machine thermal error lag phases and numerical control machine temperature lag phases as 1, 2, 3 and 4 respectively, and substituting experimental data to fit the coefficient to be solved in the formula according to a least square method so as to determine high-order multi-stage auto-regressive distributed lag models of thermal error compensation of the numerical control machine in different lag phases; substituting the experimental data into each model to obtain a residual sum of squares of each model; and substituting the residual sum of squares of each model into an akaike information criterion to determine an optimal lag phase to determine a high-order multi-stage auto-regressive distributed lag model of a thermal error of the numerical control machine. The invention discloses a modeling method of thermal error compensation of the numerical control machine, which has the advantages of convenience in application, easiness in modeling and high stability, and has higher accuracy than a traditional ADL (automatic data logger) model.

The invention provides a method for estimating weak signal separation on the basis of positioning for external radiation sources. The method includes steps of performing direct-wave and multi-path cancellation for a sample sequence x (t) of a main channel by means of NLMS (normalized least mean square) to obtain output signals y<NLMS>(t); estimating the number of source signals of the output signals y<NLMS>(t) by an AIC (Akaike information criterion) or an MDL (minimum description length) criterion, preliminarily whitening the output signals y<NLMS>(t) to obtain whitened signals z(t), and updating a separation matrix W by an ICA (independent component analysis) algorithm to obtain M channels of estimated signals y(t); respectively performing time-frequency two-dimensional correlation for the M channels of estimated signals y<1>(t), ..., y<M>(t) and reference signals s<ref>(t); and searching a peak point in absolute correlation values phi(tau, f) corresponding to the M channels of estimated signals, and determining time delay and Doppler frequency shift which correspond to the peak point as time delay and Doppler frequency shift of a target when a signal-to-noise ratio of a certain channel of signals is larger than or equal to a threshold value.

Techniques for reducing the computational complexity of conventional similarity-based approaches for temporal event clustering of digital photograph collections include one or more approaches to select boundaries based on dynamic programming and the Bayes information criterion. Each method performs competitively with conventional approaches and offer significant computational savings.

The invention discloses an output-only linear time-varying structure modal parameter identification method and belongs to the technical field of structural dynamics. Firstly, a cost function of a least squares support vector machine vector time-varying autoregressive model is deduced; secondly, a function space is built by means of a Wendland compactly supported radial basis function; a regular factor is determined through the non-parameter method based on Gamma testing, and a basis function width reduction coefficient is given on the basis of actual experiences; a time-varying autoregressive model order is determined according to the Bayesian information criterion and the Akaike information criterion; a function space order is determined according to the ratio of residual sum of squares to sequence sum of squares; finally, the matrix expression of the least squares support vector machine vector time-varying autoregressive model is solved according to the cost function, modal frequency of a system is solved according to a time freezing method, and linear time-varying structure modal parameter identification is finished. The method can improve calculation efficiency, improves system robustness, and is widely used in linear time-varying structure modal identification in structural dynamic engineering application.

The invention provides a method for predicting the key industrial electricity consumption based on an industrial condition index. The method comprises the following steps: (1) obtaining the key industrial condition index and historical electricity consumption data; (2) performing seasonal adjustment and a stationary test on the data; (3) judging whether the industrial condition index and the industrial electricity consumption have a causal relationship or not through a Granger causality test and determining an optimal lag period of the condition index; (4) creating a time sequence ARIMA (autoregressive integrated moving average) model of the key industrial electricity consumption, introducing the key industrial condition index into an original ARIMA model, and creating a regressive model; (5) on the basis of an AIC (Akaike information criterion), screening out an optimal model; (6) performing model popularization and application, and predicting the industrial electricity consumption in the future. The key industrial electricity consumption is taken as a study object, the electricity consumption and the influence of the industrial condition index on the electricity consumption are studied by introducing the industrial condition index, the key industrial electricity consumption is accurately predicted in combination with the time sequence model, and a basis is provided for development and planning of electricity industry in the future.

The invention relates to the online speaking people cluster analysis and particularly relates to the online speaking people cluster analysis based on bayesian information criterion. The method comprises steps of collecting original audio signals and dividing the original audio signals into audio segments with boundaries through the bayesian information criterion, performing audio characteristic extraction on the audio segments, clustering the segments with audio characteristics through the bayesian information criterion, and forming a plurality of clustering groups like the clustering group 1, clustering group 2....clustering group n. An online speaking people cluster analysis system comprises two modules which are an audio signal segmentation module and a segment cluster analysis module, which greatly improve the accuracy of segmentation, guarantee high efficiency of clustering, realize high efficient parallel transcription, segmentation, classification and clustering of signals of the online speaking people on a premise that the audio materials of the original speaking people are not required and realize high efficiency transcription, segmentation, classification and clustering of the signals of online speaker.

The invention provides a full-waveform laser radar data waveform decomposing method. Component number estimation is carried out through the Akaike Information Criterion (AIC), lost component detection is carried out through the Levenberg-Marquardt (LM), component number information in a back scattering echo waveform can be rapidly obtained, the algorithm is simple, and the detecting efficiency is high. Compared with an ordinary waveform decomposing method, the initialization accuracy of the model component number and the integrity of waveform decomposition are combined.

The invention discloses a data prediction method and system. The data prediction method comprises the steps of acquiring data information of a single service through a database, wherein the data information comprises a multi-dimensional characteristic variable; performing data screening on the multi-dimensional characteristic variable by combining an Akaike information criterion (AIC) value and a Bayesian information criterion (BIC) value according to a multivariate regression method, and filtering data noise; and importing the data screened characteristic variable into a machine learning model, and performing modeling analysis on the data. The accuracy of data prediction is effectively improved according to the data prediction method and system.

The invention provides a genetic locus excavation method based on a multi-target ant colony optimization algorithm. The ant colony optimization is used as a foundation design packaging type feature selection algorithm; in each type of iteration, one manual ant selects one SNPs feature subset, one feature subset and corresponding complicated property states to construct a multi-target model; a logistic regression model and a Bayesian network model are used for modeling the selected SNPs feature subset and the corresponding property states respectively; then an AIC (Akaike Information Criterion) and K2 are used as evaluation criteria of the corresponding models, and scores are used as solutions of a multi-target function; a non-dominated ranking method is adopted and all the solutions of the multi-target model in the step 2 are screened into non-dominated solutions and dominated solutions; the iteration of an pheromone matrix Tau is carried out according to advantageous and disadvantageous degrees of the solutions, and one SNPs locus subset selected by the feature selection algorithm is obtained after the iteration is finished. An endless hypothesis test is carried out on Chi-square analysis in the SNPs locus subset; finally, SNPs locus related to complicated properties is screened according to a P value set by a user.

A method and a system of predicting an electric system load based on wavelet noise reduction and empirical mode decomposition-autoregressive integrated moving average (EMD-ARIMA) are provided. The method and the system belong to a field of electric system load prediction. The method includes the following steps. Raw load data of an electric system is obtained first. Next, noise reduction processing is performed on the load data through wavelet analysis. The noise-reduced load data is further processed through an EMD method to obtain different load components. Finally, ARIMA models corresponding to the different load components are built. Further, the ARIMA models are optimized through an Akaike information criterion (AIC) and a Bayesian information criterion (BIC). The load components obtained through predicting the different ARIMA models are reconstructed to obtain a final prediction result, and accuracy of load prediction is therefore effectively improved.

The invention discloses a method for predicting service lives in consideration of long-range correlation and uncertainty of components, and belongs to the field of prediction and prediction of residual service lives in health management. The method includes steps of reading state monitoring degradation data and initializing model parameters for specified drift structures (such as exponential functions, power functions and linear functions); estimating unknown parameters of models by the aid of maximum likelihood algorithms; carrying out approximation on obtained degradation processes on the basis of weak convergence criteria and computing probability density functions of residual service lives; judging fitting effects and the prediction ability of the models by the aid of AIC (akaike information criteria) and mean-square errors. Compared with existing methods, the method has the advantages of high universality and prediction precision.

The invention modifies the signal structure of receiver to make the received signal taking the numbers of sub-carrier in each sub-channel as its periodic time in order to rearrange the signal by the receiver. After calculating covariance matrix, Jacobi eigenvalue and eigenvector, Akaike Information Criterion (AIC) is used to estimate the numbers of subscribers. After frequency deviation spectrum is figured out, through rough estimation and precision estimation, the estimated value of subscriber frequency deviation is obtained.

The invention discloses a modeling method of a support vector machine (SVM)-based software measurement instrument in a biological fermentation process. The modeling method uses an optimized SVM modelto estimate fermentation products in the biological fermentation process and comprises two steps: the SVM modeling of the biological fermentation process and the confirmation of an optimal model basedon a genetic simulated annealing algorithm and an akaike information criterion. The invention provides an optimal SVM software measurement instrument model for variables difficult in online measurement in the biological fermentation process, eliminates a phenomenon of untimely control caused by time lag of a traditional offline measuring method, lays a reliable foundation for establishing the SVMmodel in the biological fermentation software measurement and has the advantages of online measurement, automatic parameter optimization, automatic input variable selection, small sample demand and high precision.

The invention discloses a method for recognizing and modifying abnormal measurement data of a vehicle micro-mechanical gyroscope, which comprises the following steps of: firstly, selecting 400 continuous smooth data as samples in the original measurement data of the gyroscope, carrying out first difference and then zero equalization on the samples, and obtaining stable samples which are processed by the zero equalization; then, making an autocorrelation coefficient graph and a partial correlation coefficient graph of the stable samples which are processed by the zero equalization; confirming an auto-regressive moving-average model of the samples according to the distribution characteristics of autocorrelation coefficient and partial correlation coefficient of the samples as well as Akaike information criterion, and estimating the model parameters by a least-squares estimation method; after that, transforming the model of the stable samples processed by the zero equalization, and promoting the transformed model to the original measurement data; and finally, deforming the transformed model, and recognizing and modifying the abnormal measurement data of the micro-mechanical gyroscope by the deformed model.

The invention discloses a human body composition analysis method based on a genetic algorithm. The human body composition analysis method based on the genetic algorithm comprises the following steps: eight sections of human body impedance models are selectively used and an expression of each section of human body impedance is analyzed and calculated; groups of different voltage and current are set so that groups of human body impedance data models are obtained through calculation; an AIC value of each group of human body impedance data models is calculated through an akaike information criterion and combination with human body physiological parameters; a fitting model is selected; and genetic evolution is conducted on a position coefficient of the fitting model, an unknown parameter of the fitting model is determined through copy, intersection and mutation operation and a human body composition predicting formula is obtained. According to a calculation method of the eight sections of human body impedance models, theoretical reference can be provided to eight section impedance measurement technology. According to the human body composition predicting method based on the genetic algorithm, human body composition predicting accuracy can be improved and an effective detection measure is offered for human body composition research and clinical application.

The invention relates to a microseismic first arrival acquisition method based on shear wave and the Akaike information criterion, belonging to microseismic first arrival automatic extraction method research in an actual seismic exploration environment. For statistical characteristics showed by a micro seismic signal, background noise and the micro seismic signal are viewed as two stationary processes with different statistical properties, shear wave transform is used to decompose and reconstruct the micro seismic signal, at the same time, combined with an AIC model, each layer of reconstructed signal is subjected to first arrival pickup, and finally the pickup result of each layer is integrated to determine seismic two-dimensional first arrival pickup result. According to the method proposed in the invention, the preprocessing of complex background noise is not needed, time and calculation cost are saved, the energy loss of the micro seismic signal in the denoising process is avoided, the amplitude frequency characteristics of an effective signal in a micro seismic two-dimensional record is effectively retained, and an important basis is provided for oil and gas reservoir monitoring, geothermy dynamic monitoring, coal mining and dynamic engineering monitoring.

The invention provides a method for predicting network traffic and a prediction algorithm. The method, which compensates for the time-delay effect by continuously carrying out the time sliding on prediction sequences, comprises the following steps: Step 1, extracting a sample array from real traffic sequences, designating the sample array as FArray and initiating the values of three variables, M, N and m; Step 2, calculating the self-similarity index H of the sample array FArray on the basis of methods, such as periodogram, R / S analysis, wavelet analysis and the like; Step 3, estimating the order of the sample array by the AIC (Akaike Information Criterion), wherein, AIC(n,m) = lnsa + 2(n+m+1) / N (1), and determining that the order of the model is (p,q), if AIC(p,q) = min AIC(n,m); Step 4, calculating the model parameter ARMA [phi, theta], wherein, ARMA [phi, theta] = ARMA (pbest, qbest), and the calculating method comprises the following steps: (1) estimating the parameter of the autoregression part, and (2) estimating the average sliding coefficient; Step 5, calculating the coefficient vector, pij = theta1pij-1+ theta2pij-2+lambada+thetaqpij-q+phij(j>0), wherein, pi0 is equal to negative 1, and when j is larger than the sum of p and d, phij is equal to 0; and Step 6, predicting the network traffic according to the following formula: X(h) = *pij[(h)]X[t+h-j].

The invention provides a method for predicting atmospheric PM2.5 concentration based on a VARX model. The method comprises the following steps of: (1), collecting pollutant concentration data and meteorological data of a target city; (2), reading the pollutant data and the meteorological data of the target city from a database; (3), performing principal component analysis of the data, except PM2.5; (4), performing a root test of a data set; (5), constructing a VARX equation model; (6), determining a lagged order of the model by using an AIC Akaike information criterion and a SC Schwarz criterion; and (7), predicting the PM2.5 concentration by correcting and using the VARX model. The method disclosed by the invention has the characteristics of being relatively high in generalization ability, relatively high in prediction precision and the like.

The invention discloses a shutter speed model analysis method of a dynamic contrast-enhanced magnetic resonance image, wherein the method comprises the steps: obtaining a time signal of the blood vessel contrast agent concentration of a biological individual in DCE-MRI time series data; according to the time signal of the blood vessel contrast agent concentration, carrying out non-linear least square fitting on the DCE-MRI signal of each pixel with two models of SSM[full] and SSM[vas] respectively, to obtain the fitting result of the DCE-MRI signal of each pixel respectively; scoring and comparing the fitting results of the DCE-MRI signals of the pixels by a correction Akaike information criterion, and selecting an optimal model from the two models of SSM[full] and SSM[vas]; fitting according to the selected optimal model, and generating distribution diagrams of the physiological parameters K<trans>, pb, po, k[bo] and k[io]; carrying out error analysis on k[io] and k[bo] parameters, deleting the pixel results with confidence interval less than 95%, producing the final distribution diagram of the k[io] and k[bo] parameters and the final distribution diagram of the K<trans>, pb and po parameters. The method can automatically match the best SSM model of each image voxel and improve the estimation accuracy of the physiological parameters.

The invention discloses a human body composition prediction method based on AIC and improved entropy weight method. The method comprises the following steps: S1, selecting an impedance model, collecting data and constructing an original feature set F of a physiological information sample; S2, adding the original feature set F of the physiological information sample to construct a first feature parameter and a second feature parameter; S3, selecting an AIC stabilization model by use of an akaike information criterion; S4, computing the value of an AIC, and selecting a feature combination with the minimum AIC value to obtain a feature parameter matrix, analyzing the influence of each feature parameter to a fitting model, and modifying the feature parameter matrix; S5, introducing the information entropy to obtain a unified matrix, computing the entropy and the weight; and S6, solving a feature parameter matrix coefficient to obtain a human body composition fitting model. The established human body composition prediction model can improve the human body composition prediction precision and provides more effective detection means for the human body composition research and clinical application.

The invention belongs to the technical field of digital audio signal processing and discloses an automatic digital audio tampering point positioning method based on the BIC (Bayesian information criterion). The method comprises the steps as follows: performing active voice detection on a to-be-detected tampering signal to determine a silence fragment in the voice signal; sequentially extracting the Mel-scale frequency cepstral coefficient characteristic of each frame after framing of the silence fragment, and performing long window framing in time sequence; calculating the BIC value of each long-term characteristic frame; taking all crest points in a sequence constituted by BIC values of all long-term characteristic frames as suspicious tamper points, and cutting off the silence fragment front and back with the suspicious tamper points as midpoints; calculating a BIC value sequence of each cut-off window containing suspicious points. Automatic positioning of digital audio tampering points is realized; compared with a traditional tampering detection method, the method has the advantages that the calculated amount is reduced, the omission ratio of the tampering points is reduced, thethreshold selection problem is solved; the method has robustness for the condition of covering of noise with the tampering points.

The invention relates to a power demand forecast method based on a vector autoregression model. Step one, the economic data and the power demand data of a region (country, province or city) to be forecasted are collected; step two, the data are compiled and cleaned and stored in a database; step three, the time period required to be forecasted and the minimum time interval of the data are definedand logarithmic processing is performed on the corresponding data; step four, stationarity testing is performed on the four columns of processed data (fixed asset investment, total retail sales of consumer goods, total net export trade and the power demand); step five, an equation based on the vector autoregression model is constructed; step six, the lag order of the model is determined by using the AIC Akaike information criterion and the SIC Schwarz criterion; and step seven, the vector autoregression model is corrected and the power demand is forecasted by using the vector autoregression model. The method has the characteristics of high generalization capability and high forecast accuracy.

The invention discloses an esophageal squamous carcinoma radical postoperative patient prognosis prediction model construction method and device, and the method comprises the steps: obtaining clinical diagnosis and treatment data and follow-up visit survival data, carrying out multi-factor Cox regression analysis on patient characteristic variables, tumor pathology characteristic variables, treatment condition variables and test index variables according to follow-up visit survival data, carrying out variable screening by utilizing a step-by-step back algorithm and an Akaike information criterion, and carrying out variable screening on the screened candidate variables again to obtain modeling variables; and performing multi-factor Cox regression analysis on modeling variables and interaction items of every two modeling variables to construct a prognosis prediction model of a patient after the esophageal squamous carcinoma radical operation, wherein the prediction variables comprise age, gender, tumor primary position, T stage, lymph node detection number, tumor size, preoperative hemoglobin level and N stage treatment mode interaction items. According to the method, the prediction accuracy can be improved, the optimal benefit group of different treatment schemes is defined, and the prognosis evaluation precision of the esophageal squamous cell carcinoma is realized.

The invention discloses an evaluation method and system of water quality based on fish electrocardio indexes. The method comprises the following steps: receiving fish body electrocardio signals beforeand after fish contact with a solution which contains specific organic matter, and respectively extracting multiple kinds of electrocardio indexes; analyzing the relativity between environmental stress and each extracted electrocardio index according to a Pearson correlational analysis method, and screening the electrocardio indexes for evaluating the specific organic matter in the water quality;analyzing and evaluating environmental stress and linear regression models between various electrocardio indexes according to Akaike information criteria and the residual sum of squares so as to obtain the electrocardio indexes used for finally evaluating the organic matter in the water quality. Receive the fish body electrocardio signals before and after the fish contact with to-be-determined water, and respectively extracting an electrocardio index which is corresponding to a certain organic matter in determined evaluation water quality; contrast the extracted electrocardio indexes of the fish body electrocardio signals before and after the fish contact with the to-be-determined water so as to obtain electrocardio index changes for evaluating the corresponding organic matter in the to-be-determined water quality.

A load prediction method for an electric vehicle charging station based on a copula algorithm comprises the following steps of (1) classifying users; (2) classifying and fitting data; (3) extending the data; and (4) outputting a load curve of the electric vehicle charging station. According to the load prediction method for the electric vehicle charging station based on the copula algorithm, the users are firstly classified; a kernel function and a copula function are selected by use of an AIC (Akaike Information Criterion) and a BIC (Bayesian Information Criterion), data are fitted by use ofa kernel density function, extended data including a coupling relation between the data are obtained in combination with the copula algorithm, the extended data can really reflect actual charging behaviors of the users, and an obtained prediction curve better accords with the actual situation.

The invention discloses a frequency spectrum prediction method based on radio frequency machine learning model driving. The method comprises the following steps: S1, collecting frequency spectrum data, and preprocessing the collected frequency spectrum data; S2, determining an order for an autoregression model according to an akaike information criterion, and determining a step length M of input data; S3, expanding a linear combination process of the autoregression model into an M-layer network structure, and introducing new trainable parameters into the M-layer network structure to construct M layers of frequency spectrum prediction network models driven by a radio frequency machine learning model; S4, training the spectrum prediction network model by using the training set data; and S5, judging whether training is completed or not, if so, inputting test set data into the trained spectrum prediction network model, outputting a prediction result, and ending the process; and if not, adding one to the number of training iterations, and returning to step S4 until the maximum number of iterations is reached. The network is endowed with interpretability, the prediction performance is improved, and the convergence speed of the network is increased.

Scanning probe microscope (SPM) images are enhanced by enforcing one or more symmetries that can be selected based on suitable Fourier coefficient amplitude or phase angle residuals, and / or geometric Akaike information criteria, and / or cross correlation techniques. Alternatively, this selection can be based on prior knowledge of specimen characteristics. In addition, a scanning microscope point spread function is obtained based on the evaluation of a calibration image by enforcing at least one symmetry and can be applied to other image acquisitions.

The invention provides a design method of liquid ring angular accelerometer and a modeling method of a dynamic-pressure process. The modeling method of the present invention realizes high-efficiency and high-precision liquid-phase angular accelerometer dynamic-pressure process modeling based on a simplified subspace identification method, and improves the efficiency of liquid ring angular accelerometer design. The modeling method of the present invention has a complex structure and multi-variable in the dynamic-pressure process. In the subspace identification used in the present invention, orthogonal triangle decomposition is used to reduce the complexity of the algorithm. The modeling method of the present invention uses the Akaike information quantity criterion to determine the order ofthe system, and reduces the influence of sampling sample shortage and noise on system order judgment. Since the dynamic-pressure modeling process of the present invention is simplified, the design ofa liquid ring angular accelerometer based on the design method of the liquid ring angular accelerometer can improve the production efficiency of the liquid ring angular accelerometer.

The invention discloses a method for adjusting the air-conditioner energy efficiency ratio through dry-wet bulb temperature allowance grading. The method is used for performing fitting on the refrigerating capacity and energy efficiency ratio data of a room air conditioner under multiple groups of specific testing conditions, adopts an akaike information criterion (AIC) to respectively screen an optimal curved surface equation, solve an extreme value interval of two gradients and confirm a core interval (union of dry-wet bulb temperature allowance ranges in the extreme value intervals of the refrigerating capacity and energy efficiency ratio gradients ), performs three-level nonlinear division on the dry-wet bulb temperature allowance and further adjusts the air-conditioner energy efficiency ratio. The method can accurately control the performance of the room air conditioner, narrow a dry-wet bulb temperature allowance changing range and reduce the change rates of the refrigerating capacity and energy efficiency ratio.