85 results about "Data complexity" patented technology

Techniques are disclosed for selectively re-homing data processing tasks based on data complexity and candidate system performance characteristics. A server system obtains performance capabilities of a target system such that the server can decide whether to send program code and data to the target system for execution, based in part on the complexity of the data and code used in performing the request. In cases where a client system sending a request is capable of executing code and where other constraints are satisfied, e.g., bandwidth, throughput, and data volume, then the server can decide to re-home both the data and the program code to the client for execution.

A method of manufacturing semiconductor circuits or electronic packages is provided. The step of splitting up wide metal areas into metal stripes is included into the physical design step of such devices. This method does not increase the data complexity significantly and guarantees correctness of the design. Furthermore, the method allows to solve the dishing problem that is inherent to copper wiring technologies.

The invention relates to a pulse wave motion artifact removal method for parameter adaptive optimization VMD, and belongs to the technical field of biomedical signal acquisition and processing. The method comprises the following steps of by optimizing an objective function of variational mode decomposition, evaluating the variational mode decomposition result by adopting a correlation coefficientwith the fusion of the intrinsic mode correlation degree and the central frequency dispersion degree; carrying out parameter adaptive optimization variational mode decomposition on a pulse wave signalto obtain an intrinsic mode [uk] of the pulse wave, and designing a denoising rule according to the relative energy of frequency bands corresponding to different components of the signal to remove the motion artifacts of the pulse wave. Compared with the prior art, the method has the advantages that the quality of pulse wave variational mode decomposition is improved, the effect of removing the motion artifacts is improved, and the data complexity in the process of removing the motion artifacts is simplified.

The invention discloses a PCA-SIFT-based fast image splicing method. The method includes the following steps that: an improved non-maximum value suppression method is introduced in a spatial extreme point detection stage to optimize initial feature points, so that a more uniformly distributed feature point set can be obtained; 64-dimensional SIFT descriptors are extracted based on a circular domain in a descriptor construction stage, a principal component analysis (PCA) method is adopted to perform further dimensionality reduction on the descriptors; a K-D tree BBF search strategy is introduced in a feature matching stage, and a random sampling consensus (RANSAC) algorithm is adopted to eliminate mismatched points, and therefore, matching speed and matching accuracy can be improved. An image splicing experiment shows that the correct matching rate of the method of the present invention is higher than that of an SIFT algorithm, and the splicing speed of the method of the present invention is 1.6 to 2.2 times that of a conventional SIFT algorithm.

The invention discloses an EEMD (Ensemble Empirical Mode Decomposition) and combined kernel RVM (Relevance Vector Machines)-based photovoltaic power short-term prediction method. Through the adopted EEMD, a problem that the possibility of mode mixing during empirical mode decomposition is high is avoided, so that higher resolution and very strong non-linear processing capacity are provided, the data complexity is reduced, the effect is excellent, the result is accurate, the prediction accuracy is effectively improved, and the method can be applied to the data preprocessing of the photovoltaic output power well; through the adoption of an RVM method for short-term photovoltaic power prediction, the method has the advantages of being sparse in model height, less in kernel parameters to be optimized, flexible in kernel function selection, and high in model generalization ability; and by use of a combined kernel function, the prediction accuracy of photovoltaic power through a model is further improved when the weather changes suddenly, so that the universal adaptability and the generalization performance of the model are improved.

The invention discloses a K-means classifier based on a memristor array and a classification method of the K-means classifier. The dimension information of the clustering center of the K-means algorithm is used as a training weight; the tranining weight is mapped and stored in a memristor array; dimension information of a clustering center is simulated by neural network weight; calculation of theEuclidean distance is realized based on the gradual change characteristic of the memristor. Online updating of each weight of the clustering center is directly realized on a hardware circuit; data clustering of a large amount of non-normalized data on the basis of a hardware circuit is realized; the calculation complexity caused by data normalization and the circuit complexity caused by calculation weight change of an external circuit are reduced, meanwhile, the data complexity in the data distance calculation process is also reduced, the data storage time and the operation power consumption are reduced, the data interaction consumption is saved, and the calculation time is relatively short.

A computer algorithm that is able to convert a parent CGM file, which may include a plurality of disparate methods of defining objects or high-level primitives, with low-level primitives to produce a child or normalized CGM file that includes a normalized or standard way of defining these same objects. The computer algorithm iterates over the information stored in the parent CGM file to find associations in stored vectors to determine whether disparate elements, such as segments or arcs, are used to define more complex objects. It then normalizes the data into the more complex objects to produce a normalized CGM file. Therefore, the child CGM file includes standardized high-level primitives and less overall data complexity.

The invention relates to the technical field of building model processing, and discloses a building information model-based data processing method, which comprises the following steps: S1, acquiring information, namely acquiring information of a building information model, and classifying and sorting the acquired data of the building information model by a worker; S2, input and comparison: enabling the worker to input the collected data of the building information model into the receiving device, and comparing the data of the building information model input by the input device with the data of the building information model stored in the database; s3, information association. According to the invention, analysis can be carried out according to the comparison result, whether missing data is caused by screening or omission during transmission is judged, the accuracy and preciseness of information input are improved, the data complexity can be reduced through data simplification, the number of data points is reduced, meanwhile, connotation characteristics in the data are reserved, and therefore, the resource overhead during query and processing is reduced, and the query response performance is improved.

The invention discloses a wind power combined prediction method considering spatial correlation and corrected numerical weather forecast, aiming at the defects of the traditional GP based on a singlekernel function, combining a plurality of kernel functions to obtain an optimal kernel function scheme, and establishing a prediction model on the basis of the GP based on the combined kernel function. In consideration of high data complexity of multi-dimensional meteorological factors, key factors are screened through an automatic correlation judgment algorithm, and an NWP wind speed deviation correction model is established. And meanwhile, wind speed time sequences of the target wind power plant and the adjacent wind power plant are analyzed by utilizing a spatial correlation method, a timedifference corresponding to the strongest correlation is solved by utilizing a Pearson correlation coefficient method, and a spatial correlation model is established. And based on a combined weightedprediction model of the above model, a combined model weight coefficient is obtained through a Lagrange method. Therefore, the numerical weather forecast deviation correction method and the spatial correlation method are effectively combined through the combined weighted prediction model, and the prediction precision of the wind power is improved.

According to an aspect of the invention, a display apparatus includes: an image acquisition part which acquires image data of a page on which characters are written; an image reduction part which reduces the image data of the page at a predetermined reduction rate; a complexity degree calculation part which analyzes an image structure of the image data of a character in the image data, and calculates a complexity degree of the image structure; a reduction control part which determines the reduction rate for the image reduction so that the complexity degree is kept to a predetermined extent even after the reduction; an image display part which displays the image data reduced by the image reduction part so that the image better fills the display screen.

A method for determining touch input data complexity and negotiating the processing of touch inputs from a touchscreen controller to a host processor. As touch input data is received from a touchscreen, the data may be evaluated by the touchscreen controller to determine the data's complexity. When the touchscreen controller determines the received touch input data is too complex for touchscreen controller processing, the host processor may be requested to process the data. The touchscreen controller may resume processing touch input data based on low complexity or host processor availability. Complexity may be determined based on evaluations of the number and frequency of touch inputs. Additionally, the touch input data may be evaluated to determine whether it represents multi-touch input. In an aspect, the touchscreen controller may disregard received touch input data determined to be noise.

The present invention discloses a traffic flow prediction method which can reduce the data complexity and improve the prediction precision. The method provided by the embodiment of the invention comprises the following steps of: collecting and obtaining real-time traffic flow data, and employing variation mode decomposition to decompose real-time traffic flow data to a preset number of band-limiting intrinsic mode components with different feature scales; and the number of the intrinsic mode components is set to allow the sum of estimation bandwidths of all the intrinsic mode components is theminimum. An increment-type over-limit learning network is employed to establish prediction models for each intrinsic mode component, and a pre-collected historical traffic flow data is employed to perform training of the prediction models. The trained prediction models are employed to perform traffic flow prediction of each intrinsic mode component and accumulate prediction results of all the intrinsic mode components to obtain a final traffic flow prediction result.

The invention provides a website static resource processing method and device, and relates to the technical field of computers, and the method comprises the steps of obtaining a first address of a static resource address in a project file of a website, and generating a first address list; sending the first address list to a server, so that the server stores the static resources in a CDN server based on the first address list, and determines a second address corresponding to each static resource and located in the CDN server; receiving a corresponding relationship between the first address andthe second address sent by the server; and based on the corresponding relationship, replacing the first address in the project file with a second address. Therefore, the data complexity during websiteconstruction is improved, the website construction efficiency is improved, and manpower and material resources are saved.

The invention relates to a local spline embedding-based linear classification method. The method comprises the steps of inputting training data and test data; and performing supervised local spline embedding dimension reduction of the training data: constructing an intra-class graph and an inter-class graph, and selecting neighborhoods; constructing an intra-class local tangent space and an inter-class local tangent space of training data points; obtaining global low-dimensional coordinates of the training data points and obtaining corresponding optimal linear mapping; mapping the test data into a low-dimensional manifold; and classifying the test data subjected to dimension reduction by K-nearest neighbor classification to obtain category labels of the test data. According to the method, the defects that the intra-class nearest neighbor graph and the inter-class nearest neighbor graph of the data are not comprehensively analyzed and category information of the training data is not utilized in the past are overcome; and in addition, the data complexity is greatly lowered, the observability and distinguishability of the data are enhanced, and the classification accuracy of the data on a high-dimensional manifold is greatly improved.

The invention discloses an automobile appearance damage detection and recognition hardware system. The system comprises a detection box, wherein a top camera, a left camera, a right camera, a vertex angle camera, a bottom camera, a top light source, a left light source, a right light source and a bottom light source are distributed and mounted in the detection box; the image acquisition time of the top camera, the left camera, the right camera, the vertex angle camera and the bottom camera is the same. By means of the technical scheme, the illumination complexity problem in the natural environment is solved, and data complexity caused by handheld equipment is avoided, so that accuracy and stability of automobile appearance image damage detection and recognition are greatly improved.

The invention relates to a system for generating feature vectors for long texts to realize classification. The system comprises a data preprocessing module, a word vector calculation module, a high-dimensional clustering module and a long text classification module. Redundant invalid data are deleted through the data preprocessing module, so that the data complexity and the processing dimension are reduced, and the performance and the result accuracy are improved. The word vector calculation module calculates word vectors based on an improved dynamic dimension Skip-Gram algorithm, and then theword vectors are clustered through the high-dimensional clustering module. The long text classification module classifies the long texts according to a clustering result.

The invention discloses a three-phase circuit overvoltage monitoring system and a method thereof. The three-phase circuit overvoltage monitoring system comprises a front monitoring station, a front switch and multiple front monitoring equipment, wherein input ends of the front monitoring equipment are respectively connected with a three-phase bus in corresponding segments, output ends of all the front monitoring equipment are respectively connected with the front switch, an output end of the front switch is connected with the front monitoring station. According to the three-phase circuit overvoltage monitoring system and the method, the capacitance voltage division modules and the corresponding monitoring devices are employed to carry out voltage data acquisition, voltage withstand intensity is large, breakdown is not easy to occur, frequency response bandwidth can satisfy requirements of real transmission lightning wave frequency spectrums, original data has strong authenticity, thunder overvoltage data can be accurately acquired, and the monitoring effect is better; only when overvoltage occurs, the overvoltage value data and waveform data 50 cycles before and after the overvoltage value are stored and forwarded by the front monitoring station, so data complexity is greatly reduced.

The invention belongs to the technical field of quantum computing, and particularly relates to a hybrid quantum algorithm-based combinatorial optimization solving method and system and a solver architecture, aiming at a problem to be solved by a user, the problem is expressed as a combinatorial optimization normalized form expression optimization model containing a target function and constraint conditions; mapping the optimization model to Hamiltonian, and generating a quantum circuit corresponding to a target function and a mixed state Hamiltonian operator quantum circuit related to constraint conditions; by setting optimizer parameters and feasible state and angle parameters, executing a quantum circuit by using a quantum computer or a simulator; and searching an optimal approximate solution by measuring the output of the quantum circuit. According to the method, a bridge tool is arranged between a classic computer and a quantum computer, the combination optimization problem is automatically solved through quantum calculation, the data processing speed is increased, the data complexity is reduced, and the actual application scene of quantum calculation is expanded.

The invention provides a fault prediction method and system applied to a target object, and the method comprises the following steps: obtaining the criticality weight of a key part and the fault factor of the target object; comparing the real-time information of the key component with a corresponding preset threshold value according to the fault factor, and determining the current state of the keycomponent; and determining the current state of the target object by applying the criticality weight of the key component and the current state of the key component, and predicting the possibility offailure of the target object. The criticality weights of the key components are determined, so that the data complexity can be reduced, and processing is facilitated; a fault factor of the target object is determined to analyze the likelihood of a fault occurring. And by applying the fault factor, the current state of the target object can be estimated so as to predict the possibility that the target object may encounter the fault.

The invention discloses an indoor crowd counting method based on channel state information, which comprises the following steps of: filtering and smoothing original CSI (Channel State Information) signal data, minimizing noise influence, reducing data complexity through PCA (Principal Component Analysis) dimensionality reduction, converting a wireless time sequence signal into a heat map image through a sliding window, and counting the crowd in the heat map image. And a convolutional neural network (CNN) is designed to perform feature extraction on the heat map so as to map different numbers of people in corresponding environments. According to the method, the problems of high deployment cost and serious sight shielding of a traditional visual image-based method are solved, and the problem of privacy disputes is avoided. According to the method, prediction is carried out by using the CNN network based on the image after CSI conversion, on one hand, signal features are highlighted, on the other hand, correlation between the same category is better obtained, and the performance of indoor crowd counting can be effectively improved.

The invention provides a pilot working state recognition model establishing method. The method comprises the following steps of: S1, preprocessing flight data; S2, extracting features of the preprocessed flight data; S3, carrying out data dimensionality reduction on the extracted features by utilizing a Trelets algorithm; and S4, carrying out training and modeling on the dimensionality reduced data by utilizing a Gaussian process classifier optimized on the basis of cuckoo algorithm. The method has favorable time domain localization ability and is immune to noise interference so that the feature values are more cautious. According to the method, original data is reduced to lower dimensionality, so that the complexity of the dimensionality reduced data is reduced and basis is laid for further more correct classification; the parameter optimization is relatively easy and easier to converge; and global or approximate global optimization results are searched for functions as far as possible, so that the local optimization of models is avoided.