Simulation service platform data feature extraction method, device and electronic equipment

Abstract

The present application relates to a method, device and electronic device for extracting data features of a simulation service platform, and relates to the technical field of simulation data processing. The method comprises: constructing a nested lattice system, wherein the lattice points of the nested lattice system have corresponding data, constructing a lattice Gaussian distribution function based on the nested lattice system, calculating a single-point Gaussian distribution for any lattice point based on the lattice Gaussian distribution function, adding the single-point Gaussian distribution of each lattice point to obtain the overall Gaussian distribution of the nested lattice system, calculating a discrete Gaussian distribution group based on the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution, extracting key features of the data in the nested lattice system based on the discrete Gaussian distribution group, and using the key features to cluster the data in the nested lattice system, which can improve the flexibility, accuracy and efficiency of high-dimensional data feature extraction, and improving the accuracy and efficiency of the neural network model for clustering simulation data by considering the Gaussian distribution characteristics of the nested lattice by using a mathematical model.

Description

Technical Field

[0001] The present application relates to the technical field of simulation data processing, and in particular to a method, device and electronic equipment for extracting data features from a simulation service platform. Background Art

[0002] In the field of industrial design and virtual simulation services, data processing technology has always been the core driving force for innovation and optimized design. With the rapid development of information technology, especially under the wave of big data and artificial intelligence, data processing technology has undergone a significant transformation from simple to complex and from inefficient to efficient.

[0003] Feature extraction is a crucial step in data processing technology, which involves extracting features from raw data that are useful for predictive models. Currently, in the field of industrial design and virtual simulation services, large-scale, high-dimensional data usually need to be processed. When processing high-dimensional data, traditional data clustering methods often have poor clustering effects and low computational efficiency due to the problem of dimensionality disaster. In addition, the direct storage and processing of high-dimensional data requires huge storage space and computing resources, which increases the operating costs of enterprises. In the process of data clustering, the extraction of data features also faces the problems of low accuracy and large amount of calculation. How to accurately extract data features and conduct effective analysis is still a problem that needs to be solved in the current field of industrial design and virtual simulation. Summary of the invention

[0004] In order to solve or partially solve the problems existing in the related art, the present application provides a simulation service platform data feature extraction method, device and electronic device, which can process data based on nested lattices, greatly reduce the computational complexity, and improve the flexibility, accuracy and efficiency of high-dimensional data feature extraction. By adopting a mathematical model to consider the Gaussian distribution characteristics of the nested lattice, the accuracy and efficiency of the neural network model for clustering simulation data are improved.

[0005] The first aspect of the present application provides a method for extracting data features from a simulation service platform, comprising:

[0006] Constructing a nested lattice system; the lattice points of the nested lattice system have corresponding data;

[0007] constructing a lattice Gaussian distribution function based on the nested lattice system;

[0008] For any of the lattice points, calculating a single-point Gaussian distribution based on the lattice Gaussian distribution function;

[0009] Adding the single-point Gaussian distribution of each of the lattice points to obtain the overall Gaussian distribution of the nested lattice system;

[0010] Calculate a discrete Gaussian distribution group according to the single-point Gaussian distribution of each of the lattice points and the overall Gaussian distribution;

[0011] Key features of the data in the nested lattice system are extracted based on the discrete Gaussian distribution group, and the key features are used to cluster the data in the nested lattice system.

[0012] Preferably, the steps before setting the Gaussian distribution function based on the nested lattice system include:

[0013] Setting the theta series for the nested lattice system;

[0014] Traversing the lattice points of the nested lattice system and calculating theta-order values ​​of the nested lattice system;

[0015] The parameters of the theta series are adjusted according to the theta series value until a first preset condition is met.

[0016] Preferably, the setting of a Gaussian distribution function based on the nested lattice system comprises:

[0017] When the first preset condition is met, determining the variance and the center of the Gaussian distribution;

[0018] Constructing the lattice Gaussian distribution function according to the variance and the Gaussian distribution center;

[0019] Preferably, the calculating of a discrete Gaussian distribution group according to the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution comprises:

[0020] Set up the initial data set;

[0021] For each of the lattice points, dividing the single-point Gaussian distribution of the lattice point by the overall Gaussian distribution to obtain a discrete Gaussian distribution of the lattice point;

[0022] The discrete Gaussian distribution of the lattice points is placed at corresponding positions of the initial array to generate the discrete Gaussian distribution group.

[0023] Preferably, said constructing a nested lattice system comprises:

[0024] Obtaining preset basis vectors, and constructing a generator matrix based on the preset basis vectors;

[0025] generating a first lattice according to the generator matrix;

[0026] Adjusting the generator matrix, and generating a second lattice based on the adjusted generator matrix; the second lattice is a subset of the first lattice;

[0027] The nested lattice system is generated according to the first lattice and the second lattice.

[0028] Preferably, the method further comprises:

[0029] Get input data;

[0030] Determine the lattice point with the shortest distance from the input data based on a preset quantization function;

[0031] An index corresponding to the input data and the lattice point is established, and the input data is mapped to the lattice point.

[0032] Preferably, the method further comprises:

[0033] Dividing the first lattice and the second lattice into a plurality of Voronoi units according to the generator matrix and the adjusted generator matrix respectively;

[0034] A preset number of sample points are randomly generated in a preset unit hypercube;

[0035] Determining the lattice points to which the sample points are mapped in different lattices and the Voronoi units of the sample points in different lattices by using the preset quantization function;

[0036] The number of sample points in the Voronoi units in different lattices is counted, and the ratio of the number of sample points in each Voronoi unit to the preset number of sample points is calculated respectively, and the ratio is determined as the volume of the Voronoi unit of the corresponding lattice, and the volumes of the Voronoi units of the two lattices are divided to obtain the nesting ratio of the nested lattice system.

[0037] Preferably, the method further comprises:

[0038] In any lattice, for each of the sample points, calculate the square value of the distance from each of the sample points to the mapped lattice point;

[0039] Accumulate the distance square value corresponding to each of the samples and divide it by the preset number to generate the second-order moment of the Voronoi unit;

[0040] A variance of the Voronoi cell is generated based on the second order moment, the dimension of the basis vectors, and the volume of the Voronoi cell.

[0041] A second aspect of the present application provides a simulation service platform data feature extraction device, comprising:

[0042] A construction module, used to construct a nested lattice system; the lattice points of the nested lattice system have corresponding data;

[0043] A function module, used for constructing a lattice Gaussian distribution function based on the nested lattice system;

[0044] A first calculation module, configured to calculate a single-point Gaussian distribution for any of the lattice points based on the lattice Gaussian distribution function;

[0045] A second calculation module, configured to add the single-point Gaussian distribution of each of the lattice points to obtain an overall Gaussian distribution of the nested lattice system;

[0046] A third calculation module, used for calculating a discrete Gaussian distribution group according to the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution;

[0047] An extraction module is used to extract key features of the data in the nested lattice system based on the discrete Gaussian distribution group, and the key features are used to cluster the data in the nested lattice system.

[0048] A third aspect of the present application provides an electronic device, including:

[0049] Processor; and

[0050] The memory stores executable codes thereon, and when the executable codes are executed by the processor, the processor is caused to execute the method as described above.

[0051] The technical solution provided by the present application may include the following beneficial effects: An embodiment of the present application discloses a method for extracting data features from a simulation service platform, including constructing a nested lattice system; the lattice points of the nested lattice system have corresponding data, a lattice Gaussian distribution function is constructed based on the nested lattice system, for any lattice point, a single-point Gaussian distribution is calculated based on the lattice Gaussian distribution function, the single-point Gaussian distribution of each lattice point is added to obtain the overall Gaussian distribution of the nested lattice system, a discrete Gaussian distribution group is calculated based on the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution, key features of the data in the nested lattice system are extracted based on the discrete Gaussian distribution group, the key features are used to cluster the data in the nested lattice system, the flexibility, accuracy and efficiency of high-dimensional data feature extraction can be improved, and the accuracy and efficiency of the neural network model for clustering simulation data can be improved by adopting a mathematical model to consider the Gaussian distribution characteristics of the nested lattice.

[0052] The technical solution of the present application can also: provide an accurate distance measurement basis for subsequent data clustering analysis by simulating the performance and characteristics of the Gaussian distribution on the lattice. During the clustering process, the similarities and differences between the points can be evaluated based on the discrete Gaussian distribution density values ​​on different lattice points, thereby achieving effective data clustering and feature extraction. In addition, the calculation results of the theta series can also be used to evaluate the impact of different lattice structures on the characteristics of the Gaussian distribution, providing strong support for optimizing the lattice design and improving the clustering effect.

[0053] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The above and other objects, features and advantages of the present application will become more apparent by describing in more detail the exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.

[0055] Figure 1 It is a flow chart of a method for extracting data features from a simulation service platform according to an embodiment of the present application;

[0056] Figure 2 It is another flow chart of the method for extracting data features from a simulation service platform shown in an embodiment of the present application;

[0057] Figure 3 is a schematic flow chart of a nested lattice system construction method shown in an embodiment of the present application;

[0058] Figure 4 It is a structural diagram of a simulation service platform data feature extraction device shown in an embodiment of the present application;

[0059] Figure 5 It is a schematic diagram of the structure of an electronic device shown in an embodiment of the present application. DETAILED DESCRIPTION

[0060] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0061] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0062] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0063] Feature extraction is a crucial step in data processing technology, which involves extracting features from raw data that are useful for predictive models. Currently, in the field of industrial design and virtual simulation services, large-scale, high-dimensional data usually need to be processed. When processing high-dimensional data, traditional data clustering methods often have poor clustering effects and low computational efficiency due to the problem of dimensionality disaster. In addition, the direct storage and processing of high-dimensional data requires huge storage space and computing resources, which increases the operating costs of enterprises. In the process of data clustering, the extraction of data features also faces the problems of low accuracy and large amount of calculation. How to accurately extract data features and conduct effective analysis is still a problem that needs to be solved in the current field of industrial design and virtual simulation.

[0064] In response to the above problems, an embodiment of the present application provides a simulation service platform data feature extraction method, which can improve the flexibility, accuracy and efficiency of high-dimensional data feature extraction, and improve the accuracy and efficiency of data clustering by considering the Gaussian distribution characteristics of nested lattices.

[0065] The technical solution of the embodiments of the present application is described in detail below with reference to the accompanying drawings.

[0066] Figure 1 It is a flow chart of the simulation service platform data feature extraction method shown in the embodiment of the present application.

[0067] See also Figure 1 , the method comprising:

[0068] Step 101, constructing a nested lattice system; the lattice points of the nested lattice system have corresponding data.

[0069] In the embodiment of the present application, a nested lattice system is introduced to process data, and the nested lattice system represents and processes data at different precision levels, thereby improving the flexibility and efficiency of data processing. The nested lattice system is composed of a number of lattice points, and the lattice points have corresponding data.

[0070] Step 102: construct a lattice Gaussian distribution function based on the nested lattice system.

[0071] A lattice Gaussian distribution function can be constructed based on the nested lattice system. In nature, Gaussian distribution is a common distribution form of data, characterized in that the data is high in the middle and low on both sides, with most of the data concentrated near the average value and extreme values ​​distributed at both ends. The lattice Gaussian distribution function is constructed in the embodiment of the present application to simulate the form of the Gaussian distribution of the data, which is closer to the characteristics of the data itself, so that the characteristics of the data can be extracted more accurately, thereby achieving a better clustering effect.

[0072] Step 103: For any lattice point, calculate a single-point Gaussian distribution based on the lattice Gaussian distribution function.

[0073] For any lattice point on the nested lattice system, the single-point Gaussian distribution of the lattice point can be calculated according to the lattice Gaussian distribution function.

[0074] Step 104 , adding the single-point Gaussian distribution of each lattice point to obtain the overall Gaussian distribution of the nested lattice system.

[0075] Traverse each lattice point on the nested lattice system, calculate the single-point Gaussian distribution of each lattice point, add up the single-point Gaussian distributions of all lattice points, and obtain the overall Gaussian distribution of the nested lattice system.

[0076] Step 105 , calculating a discrete Gaussian distribution group according to the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution.

[0077] The discrete Gaussian distribution group can be calculated based on the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution, achieving an accurate description of the Gaussian distribution density at a specific lattice point on the lattice.

[0078] Step 106 , extracting key features of the data in the nested lattice system based on the discrete Gaussian distribution group, where the key features are used to cluster the data in the nested lattice system.

[0079] Based on the discrete Gaussian distribution group, the key features of the data in the nested lattice system, such as distribution density, shape parameters, etc., can be extracted. Based on these key features, the neural network model can cluster the data in the nested lattice system. At the same time, based on the key features, the neural network model can also better understand the intrinsic structure of the data.

[0080] The technical solution of the embodiments of the present application can be applied to various scenarios that require continuous data clustering, such as image processing, signal processing, machine learning and other fields. By accurately simulating and analyzing the Gaussian distribution characteristics on the lattice, more efficient and accurate data clustering and feature extraction can be achieved.

[0081] The embodiment of the present application discloses a method for extracting data features of a simulation service platform, including constructing a nested lattice system, wherein the lattice points of the nested lattice system have corresponding data, constructing a lattice Gaussian distribution function based on the nested lattice system, calculating a single-point Gaussian distribution for any lattice point based on the lattice Gaussian distribution function, adding the single-point Gaussian distribution of each lattice point to obtain an overall Gaussian distribution of the nested lattice system, calculating a discrete Gaussian distribution group based on the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution, extracting key features of the data in the nested lattice system based on the discrete Gaussian distribution group, and the key features are used to cluster the data in the nested lattice system, which can improve the flexibility, accuracy and efficiency of high-dimensional data feature extraction, and improve the accuracy and efficiency of neural network model clustering of simulation data by taking into account the Gaussian distribution characteristics of the nested lattice by adopting a mathematical model.

[0082] Figure 2 It is a flow chart of the simulation service platform data feature extraction method shown in the embodiment of the present application.

[0083] See also Figure 2 , the method comprising:

[0084] Step 201, constructing a nested lattice system; the lattice points of the nested lattice system have corresponding data.

[0085] In the embodiment of the present application, a nested lattice system is introduced to process data, and the nested lattice system represents and processes data at different precision levels, thereby improving the flexibility and efficiency of data processing. The nested lattice system is composed of a number of lattice points, and the lattice points have corresponding data.

[0086] Step 202, setting the theta series for the nested lattice system.

[0087] In the embodiment of the present application, the theta series is an important parameter for measuring the density of Gaussian distribution on the lattice, and its calculation process directly affects the accuracy and efficiency of subsequent data feature extraction. The nested lattice system Λ can be any n-dimensional discrete point set, such as a common two-dimensional square lattice or a three-dimensional cubic lattice. Define the theta series Θ_Λ(q) of the nested lattice system, where q is a complex number e^jπz (z is a pure imaginary number, and its imaginary part I(z)>0). By introducing the variable τ=I(z), the theta series is converted to the real number domain, that is, Θ_Λ(τ). This conversion makes the calculation of the theta series more intuitive and convenient for numerical implementation. Θ_Λ(τ) can be expressed by formula (1):

[0088] (1)

[0089] Where τ is a real number parameter greater than 0, λ is an arbitrary point in the nested lattice system Λ,

[0090] ‖λ‖ represents the Euclidean norm of λ.

[0091] Step 203, traverse the lattice points of the nested lattice system and calculate the theta level value of the nested lattice system.

[0092] To realize the calculation of theta series values, a preset calculation function is set for theta series. The preset calculation function accepts the lattice point set Lambda and the parameter As input, the theta value is obtained by traversing each lattice point in Lambda and calculating the corresponding exponential terms and then summing them. This process not only ensures the accuracy of the calculation, but also improves the calculation efficiency by optimizing the loop structure.

[0093] Step 204, adjusting the parameters of the theta series according to the theta series value until a first preset condition is met.

[0094] The theta-level value reflects the density of the Gaussian distribution of lattice points on the nested lattice system. The parameter τ of the theta series is adjusted according to the theta-level value. By changing the value of parameter τ, the trend of the theta-level value changing with the lattice structure and τ parameter can be observed until the first preset condition is met. The first preset condition can be that the theta-level value reaches the required accuracy or reaches the boundary of the nested lattice system.

[0095] By calculating the theta series, the density and distribution of the Gaussian distribution on the lattice can be quantified, providing a basis for subsequent Gaussian distribution feature extraction.

[0096] Step 205: When the first preset condition is met, determine the variance and the center of the Gaussian distribution.

[0097] When the first preset condition is met, the variance σ^2 and the Gaussian distribution center c (c∈R^n) may be determined for the Gaussian distribution function.

[0098] Step 206: construct a lattice Gaussian distribution function according to the variance and the Gaussian distribution center.

[0099] The lattice Gaussian distribution function f_(σ,c)(x) can be constructed based on the variance σ^2 and the Gaussian distribution center c. The lattice Gaussian distribution function f_(σ,c)(x) can be expressed by formula (2):

[0100] (2)

[0101] This function is applicable to x∈R^n (n is the data dimension) and follows the basic form of Gaussian distribution.

[0102] Step 207 : for any lattice point, calculate a single-point Gaussian distribution based on the lattice Gaussian distribution function.

[0103] For any lattice point λ, the Euclidean distance dist = ‖λ- c‖ between the lattice point λ and the center c of the Gaussian distribution is first calculated, and then the single-point Gaussian distribution f_(σ,Λ)(c) is calculated based on the lattice Gaussian distribution function. The single-point Gaussian distribution f_(σ,Λ)(c) can be calculated using formula (3):

[0104] f_(σ,c)(λ) = (1 / (σ√(2π))) * e^ (-0.5 * (dist / σ) ^ 2) (3)

[0105] Among them, f_(σ,c)(λ) is centered at λ and σ is the standard deviation.

[0106] Step 208 , adding the single-point Gaussian distribution of each lattice point to obtain the overall Gaussian distribution of the nested lattice system.

[0107] Traverse each lattice point in the nested lattice system, calculate the single-point Gaussian distribution of each lattice point, add up the single-point Gaussian distributions of all lattice points, and obtain the overall Gaussian distribution of the nested lattice system (f_(σ,Λ)(c).

[0108] Step 209 , calculating a discrete Gaussian distribution group according to the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution.

[0109] The discrete Gaussian distribution group can be calculated based on the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution, achieving an accurate description of the Gaussian distribution density at a specific lattice point on the lattice.

[0110] In an optional embodiment of the present application, step 209 includes:

[0111] Sub-step 2091, setting the initial data group.

[0112] An initial data sub-D may be set first, the size of which is the same as the lattice point set Lambda.

[0113] Sub-step 2092 , for each lattice point, divide the single-point Gaussian distribution of the lattice point by the overall Gaussian distribution to obtain a discrete Gaussian distribution of the lattice point.

[0114] Traverse each lattice point λ in the lattice point set Lambda, divide the single-point Gaussian distribution of the lattice point by the overall Gaussian distribution, and obtain the discrete Gaussian distribution D_(Λ,σ,c)(λ) of the lattice point. The discrete Gaussian distribution D_(Λ,σ,c)(λ) can be calculated using formula (4):

[0115] (4)

[0116] Among them, this step needs to ensure that the denominator is not zero, that is, the total Gaussian distribution contribution at the center c of the Gaussian distribution is not zero.

[0117] Sub-step 2093, placing the discrete Gaussian distribution of the lattice points at the corresponding positions of the initial array to generate a discrete Gaussian distribution group.

[0118] The discrete Gaussian distribution D_(Λ,σ,c)(λ) of the lattice point λ is stored in the corresponding position of the initial array D to generate a discrete Gaussian distribution group.

[0119] Step 2010: extract key features of the data in the nested lattice system based on the discrete Gaussian distribution group, and the key features are used to cluster the data in the nested lattice system.

[0120] Based on the discrete Gaussian distribution group, the key features of the data in the nested lattice system, such as distribution density, shape parameters, etc., can be extracted. Based on these key features, the data in the nested lattice system can be clustered. At the same time, based on the key features, the inherent structure of the data can be better understood.

[0121] The embodiment of the present application discloses a method for extracting data features from a simulation service platform, including constructing a nested lattice system; the lattice points of the nested lattice system have corresponding data, setting a theta-level value for the nested lattice system, traversing the lattice points of the nested lattice system, calculating the theta-level value of the nested lattice system, adjusting the parameters of the theta-level according to the theta-level value until a first preset condition is met, determining the variance and the center of the Gaussian distribution when the first preset condition is met, constructing a lattice Gaussian distribution function according to the variance and the center of the Gaussian distribution, calculating a single-point Gaussian distribution for any lattice point based on the lattice Gaussian distribution function, adding the single-point Gaussian distribution of each lattice point to obtain an overall Gaussian distribution of the nested lattice system, and obtaining the overall Gaussian distribution of the nested lattice system. The discrete Gaussian distribution group is calculated according to the single-point Gaussian distribution and the overall Gaussian distribution of each lattice point. The key features of the data in the nested lattice system are extracted based on the discrete Gaussian distribution group. The key features are used to cluster the data in the nested lattice system. By simulating the performance and characteristics of the Gaussian distribution on the lattice, it can provide an accurate distance measurement basis for subsequent data clustering analysis. In the clustering process, the similarities and differences between the points can be evaluated according to the discrete Gaussian distribution density values ​​on different lattice points, thereby achieving effective data clustering and feature extraction. In addition, the calculation results of the theta series can also be used to evaluate the impact of different lattice structures on the Gaussian distribution characteristics, providing strong support for optimizing the lattice design and improving the clustering effect.

[0122] Figure 3 It is a flow chart of the nested lattice system construction method shown in the embodiment of the present application.

[0123] See also Figure 3 , the method comprising:

[0124] Step 301, obtain preset basis vectors, and construct a generator matrix based on the preset basis vectors.

[0125] Before constructing a nested lattice system, you can first analyze the statistical characteristics and distribution patterns of the data set, such as the mean, covariance matrix, etc. These data will guide the selection of basis vectors to ensure that the basis vectors can better reflect the intrinsic structure of the data. Based on the characteristics of the data, select a set of n basis vectors {m_1, m_2, ..., m_n}, where n is the data dimension. These basis vectors should be as orthogonal as possible and have uniform length to reduce redundancy and errors in data representation. Methods such as principal component analysis (PCA) or independent component analysis (ICA) can be used to assist in the selection of basis vectors. Arrange the selected basis vectors in columns and construct an nxn generating matrix M = [m_1, m_2, ..., m_n]. Ensure that the column vectors of M are linearly independent to form an effective n-dimensional spatial basis.

[0126] Step 302: Generate a first lattice according to a generator matrix.

[0127] Using the generating matrix M, according to the definition of the lattice Λ={λ= M•i, i∈Zn}, where Zn represents the n-dimensional integer space, the n-dimensional first lattice Λ_F is generated. The first lattice Λ_F is a fine lattice. The first lattice Λ_F can cover the entire n-dimensional data space, and its distribution and density in the space can be adjusted by selecting basis vectors. Through linear combinations of integer multiples of M, the points in the entire n-dimensional data space can be traversed.

[0128] Step 303, adjusting the generator matrix, and generating a second lattice based on the adjusted generator matrix; the second lattice is a subset of the first lattice.

[0129] Adjust the generation matrix M, generate k (k is a positive integer) based on the generation matrix M, record it as kM, and generate the second lattice Λ_C according to kM using the same method as in step 302. The second lattice Λ_C is a coarse lattice, and the second lattice Λ_C is a subset of the first lattice Λ_F.

[0130] Step 304 : generating a nested lattice system according to the first lattice and the second lattice.

[0131] A nested lattice system Λ(Λ_F, Λ_C) is generated based on the first lattice Λ_F and the second lattice Λ_C. This nested relationship allows the data to be represented and processed at different levels of accuracy, which improves the flexibility and efficiency of data processing. In one example, the structure of the nested lattice system Λ(Λ_F, Λ_C) in a two-dimensional or three-dimensional subspace can also be visualized (if n is large, a representative subspace can be selected for visualization) to check whether the nested lattice system covers the data space evenly and make adjustments as needed.

[0132] In an optional embodiment of the present application, the method further includes:

[0133] Step S11, obtaining input data.

[0134] Get an arbitrary input data point x.

[0135] Step S12: determining the lattice point with the shortest distance to the input data based on a preset quantization function.

[0136] A preset quantization function Q_Λ(x) can be set for the nested lattice system, which receives a data point x as input. It traverses the points in the nested lattice system Λ(Λ_F, Λ_C) and finds the lattice point λ with the smallest Euclidean distance to the data point x, that is, Q_Λ(x) = argmin_(λ∈Λ) |x-λ|.

[0137] Step S13, establishing an index corresponding to the input data and the lattice point, and mapping the input data to the lattice point.

[0138] An index between the input data point x and the determined lattice point λ is established, and the input data point x is mapped to the lattice point λ. In one example, the performance of the preset quantization function can also be optimized, for example, by constructing a search tree (such as a kd tree or a ball tree) to accelerate the search process of the nearest neighbor point and improve the quantization efficiency. By presetting the quantization function, the data points are discretized and compressed, effectively reducing the burden of data storage and processing.

[0139] In an optional embodiment of the present application, the method further includes:

[0140] Step S21 , dividing the first lattice and the second lattice into a plurality of Voronoi units according to the generator matrix and the adjusted generator matrix.

[0141] The first lattice Λ_F and the second lattice Λ_C can be divided into a plurality of Voronoi cells according to the generator matrix and the adjusted generator matrix, respectively. In one example, for any lattice, by presetting the quantization function Q_Λ, a basic Voronoi cell V of the lattice Λ can be defined, and the cell contains data points quantized to the origin 0, that is, V = {x: Q_Λ(x) = 0}.

[0142] Step S22: randomly generate a preset number of sample points in a preset unit hypercube.

[0143] A large number of preset sample points, such as 10,000 sample points, are randomly generated in the preset unit hypercube [0,1]^n. These sample points should be evenly distributed in the entire preset unit hypercube.

[0144] Step S23, determining the lattice points to which the sample points are mapped in different lattices and the Voronoi units of the sample points in different lattices by using a preset quantization function.

[0145] For each sample point, the preset quantization function Q_Lambda is used to find the corresponding nearest lattice point, and the Voronoi unit to which the sample point belongs is recorded.

[0146] Step S24, counting the number of sample points in the Voronoi units in different lattices, respectively calculating the ratio of the number of sample points in each Voronoi unit to the preset number of sample points, determining the ratio as the volume of the Voronoi unit of the corresponding lattice, dividing the volumes of the Voronoi units of the two lattices, and obtaining the nesting ratio of the nested lattice system.

[0147] In a lattice, the number of sample points falling in each Voronoi cell can be counted, and the proportion of the total number of sample points in all Voronoi cells to the total number of sample points can be calculated. Since the volume of the unit hypercube is 1, this ratio is approximately the Voronoi cell volume. Using this method, the Voronoi cell volumes V_F and V_C of the first lattice Λ_F and the second lattice Λ_C can be calculated respectively, and the volumes V_F and V_C of the first lattice Λ_F and the second lattice Λ_C are divided to obtain the nesting ratio of the nested lattice system. The nesting ratio reflects the trade-off between the accuracy and efficiency of data representation under different lattice structures. If the Voronoi cell volume of the second lattice Λ_C is larger, the smaller the nesting ratio, the lower the accuracy of the data, but the higher the efficiency. By adjusting the nesting ratio, the accuracy and computational complexity of data clustering can be optimized. In practical applications, a suitable nesting ratio can be selected according to data characteristics and processing requirements to achieve efficient and accurate data clustering.

[0148] In an optional embodiment of the present application, the method further includes:

[0149] Step S31 : in any lattice, for each sample point, calculate the square value of the distance from each sample point to the mapped lattice point.

[0150] In any lattice, for each sample point, a squared distance value from each sample point to the mapped lattice point is calculated, wherein the squared distance value may be a squared Euclidean distance.

[0151] Step S32, summing up the squared distance values ​​corresponding to each sample and dividing by a preset number to generate the second-order moment of the Voronoi unit.

[0152] The squared distance value of each sample is accumulated and then divided by the total number of sample points to obtain the average squared distance value, which is the second-order moment of the Voronoi cell. The volume V and second-order moment of the Voronoi cell are important parameters for measuring lattice performance and are used to evaluate the density of lattice point distribution.

[0153] Step S33, generating the variance of the Voronoi unit based on the second-order moment, the dimension of the basis vector and the volume of the Voronoi unit.

[0154] Based on the volume and second-order moment of the Voronoi cell and the known dimension n of the basis vector, the variance σ^2 of the Voronoi cell is calculated. The variance of the Voronoi cell is calculated using the following formula (5):

[0155] (5)

[0156] Through the Monte Carlo method, random sampling and statistical analysis are used to estimate the key properties of the Voronoi unit, which greatly reduces the computational complexity and improves the computational efficiency. It does not depend on the specific Voronoi unit shape and position, and only needs to define the basis vector and lattice structure, so it has wide applicability. In the embodiment of the present application, the accuracy of the calculation can also be improved by increasing the number of sample points, and the elements of the generator matrix M can also be adjusted to observe the influence on the Voronoi unit characteristics and nesting ratio to optimize the lattice structure. In practical applications, the technical solution of the embodiment of the present application can also be further adjusted and optimized according to the specific characteristics of the data, such as distribution density, correlation, etc.

[0157] An embodiment of the present application discloses a method for constructing a nested lattice system, including: obtaining preset basis vectors, constructing a generator matrix based on the preset basis vectors, generating a first lattice according to the generator matrix, adjusting the generator matrix, and generating a second lattice based on the adjusted generator matrix; the second lattice is a subset of the first lattice, and a nested lattice system is generated based on the first lattice and the second lattice. The parameters such as the generator matrix and the basis vector can be adjusted according to actual needs to adapt to different application scenarios and data characteristics, and the flexibility is higher. By adopting a nested lattice system, data can be represented and processed at different precision levels, thereby improving the flexibility and efficiency of data processing.

[0158] Corresponding to the aforementioned application function implementation method embodiment, the present application also provides a simulation service platform data feature extraction device, electronic device and corresponding embodiments.

[0159] Figure 4 It is a structural diagram of a simulation service platform data feature extraction device shown in an embodiment of the present application.

[0160] See also Figure 4 , the data feature extraction device 400 comprises:

[0161] A construction module 401 is used to construct a nested lattice system; the lattice points of the nested lattice system have corresponding data;

[0162] Function module 402, used to construct a lattice Gaussian distribution function based on the nested lattice system;

[0163] A first calculation module 403 is used to calculate a single-point Gaussian distribution based on a lattice Gaussian distribution function for any lattice point;

[0164] A second calculation module 404 is used to add the single-point Gaussian distribution of each lattice point to obtain the overall Gaussian distribution of the nested lattice system;

[0165] A third calculation module 405 is used to calculate a discrete Gaussian distribution group according to the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution;

[0166] The extraction module 406 is used to extract key features of the data in the nested lattice system based on the discrete Gaussian distribution group, and the key features are used to cluster the data in the nested lattice system.

[0167] In an optional embodiment of the present application, the device further includes:

[0168] The series module is used to set theta series for nested lattice systems;

[0169] The level value module is used to traverse the lattice points of the nested lattice system and calculate the theta level value of the nested lattice system;

[0170] The adjustment module is used to adjust the parameters of the theta series according to the theta series value until a first preset condition is met.

[0171] In an optional embodiment of the present application, the function module 402:

[0172] A determination submodule, used to determine the variance and the center of the Gaussian distribution when the first preset condition is met;

[0173] Function submodule, used to construct the lattice Gaussian distribution function based on the variance and the center of the Gaussian distribution;

[0174] In an optional embodiment of the present application, the third calculation module 405 includes:

[0175] The initial submodule is used to set the initial data set;

[0176] A distribution submodule, for dividing the single-point Gaussian distribution of the lattice point by the overall Gaussian distribution for each lattice point, to obtain a discrete Gaussian distribution of the lattice point;

[0177] The distribution group submodule is used to place the discrete Gaussian distribution of the lattice points at the corresponding positions of the initial array to generate a discrete Gaussian distribution group.

[0178] In an optional embodiment of the present application, the construction module 401 includes:

[0179] A basis vector quantum module is used to obtain preset basis vectors and construct a generator matrix based on the preset basis vectors;

[0180] A first lattice submodule, used to generate a first lattice according to a generation matrix;

[0181] A second lattice submodule, used to adjust the generator matrix and generate a second lattice based on the adjusted generator matrix; the second lattice is a subset of the first lattice;

[0182] The nested lattice submodule is used to generate a nested lattice system according to the first lattice and the second lattice.

[0183] In an optional embodiment of the present application, the device further includes:

[0184] Input module, used to obtain input data;

[0185] A quantization module, used for determining a lattice point with a minimum distance from input data based on a preset quantization function;

[0186] The mapping module is used to establish the index corresponding to the input data and the lattice points, and map the input data to the lattice points.

[0187] In an optional embodiment of the present application, the device further includes:

[0188] A partitioning module, used for partitioning the first lattice and the second lattice into a plurality of Voronoi units according to the generator matrix and the adjusted generator matrix;

[0189] A random module, used to randomly generate a preset number of sample points in a preset unit hypercube;

[0190] A determination module, used for determining the lattice points mapped by the sample points in different lattices and the Voronoi units of the sample points in different lattices through a preset quantization function;

[0191] The statistical module is used to count the number of sample points in the Voronoi unit in different lattices, calculate the ratio of the number of sample points in each Voronoi unit to the preset number of sample points, determine the ratio as the volume of the Voronoi unit of the corresponding lattice, divide the volumes of the Voronoi units of the two lattices, and obtain the nesting ratio of the nested lattice system.

[0192] In an optional embodiment of the present application, the device further includes:

[0193] A distance square value module, used for calculating the distance square value from each sample point to the mapped lattice point for each sample point in any lattice;

[0194] The second-order moment module is used to accumulate the squared distance values ​​corresponding to each sample and divide it by a preset number to generate the second-order moment of the Voronoi unit;

[0195] A variance module is used to generate the variance of the Voronoi cell based on the second-order moment, the dimension of the basis vectors, and the volume of the Voronoi cell.

[0196] The embodiment of the present application discloses a data feature extraction device for a simulation service platform, including constructing a nested lattice system; the lattice points of the nested lattice system have corresponding data, constructing a lattice Gaussian distribution function based on the nested lattice system, calculating a single-point Gaussian distribution for any lattice point based on the lattice Gaussian distribution function, adding the single-point Gaussian distribution of each lattice point to obtain an overall Gaussian distribution of the nested lattice system, calculating a discrete Gaussian distribution group based on the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution, extracting key features of the data in the nested lattice system based on the discrete Gaussian distribution group, the key features are used to cluster the data in the nested lattice system, which can improve the flexibility, accuracy and efficiency of high-dimensional data feature extraction, and improve the accuracy and efficiency of data clustering by considering the Gaussian distribution characteristics of the nested lattice.

[0197] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated again here.

[0198] Figure 5 It is a schematic diagram of the structure of an electronic device shown in an embodiment of the present application.

[0199] See also Figure 5 , the electronic device 500 includes a memory 510 and a processor 520 .

[0200] The processor 520 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0201] The memory 510 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage devices. Among them, ROM can store static data or instructions required by the processor 520 or other modules of the computer. The permanent storage device may be a readable and writable storage device. The permanent storage device may be a non-volatile storage device that does not lose the stored instructions and data even after the computer is powered off. In some embodiments, the permanent storage device uses a large-capacity storage device (such as a magnetic or optical disk, flash memory) as a permanent storage device. In some other embodiments, the permanent storage device may be a removable storage device (such as a floppy disk, optical drive). The system memory may be a readable and writable storage device or a volatile readable and writable storage device, such as a dynamic random access memory. The system memory may store some or all instructions and data required by the processor at runtime. In addition, the memory 510 may include any combination of computer-readable storage media, including various types of semiconductor storage chips (such as DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and disks and / or optical disks may also be used. In some embodiments, the memory 510 may include a readable and / or writable removable storage device, such as a laser disc (CD), a read-only digital versatile disc (such as a DVD-ROM, a double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (such as an SD card, a mini SD card, a Micro-SD card, etc.), a magnetic floppy disk, etc. Computer-readable storage media do not include carrier waves and transient electronic signals transmitted wirelessly or wired.

[0202] The memory 510 stores executable codes, and when the executable codes are processed by the processor 520 , the processor 520 can execute part or all of the methods described above.

[0203] In addition, the method according to the present application may also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing some or all of the steps in the above method of the present application.

[0204] Alternatively, the present application can also be implemented as a computer-readable storage medium (or non-transitory machine-readable storage medium or machine-readable storage medium) on which executable code (or computer program or computer instruction code) is stored. When the executable code (or computer program or computer instruction code) is executed by a processor of an electronic device (or server, etc.), the processor executes part or all of the steps of the above-mentioned method according to the present application.

[0205] The embodiments of the present application have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method for extracting data features from a simulation service platform, characterized in that: The method comprises: Constructing a nested lattice system; the lattice points of the nested lattice system have corresponding data; constructing a lattice Gaussian distribution function based on the nested lattice system; For any of the lattice points, calculating a single-point Gaussian distribution based on the lattice Gaussian distribution function; Adding the single-point Gaussian distribution of each of the lattice points to obtain the overall Gaussian distribution of the nested lattice system; Calculate a discrete Gaussian distribution group according to the single-point Gaussian distribution of each of the lattice points and the overall Gaussian distribution; extracting key features of the data in the nested lattice system based on the discrete Gaussian distribution group, wherein the key features are used to cluster the data in the nested lattice system; Wherein, constructing the nested lattice system includes: obtaining preset basis vectors, constructing a generator matrix based on the preset basis vectors; generating a first lattice according to the generator matrix; adjusting the generator matrix, generating a second lattice based on the adjusted generator matrix; the second lattice is a subset of the first lattice; generating the nested lattice system according to the first lattice and the second lattice; The method further includes: acquiring input data; determining a lattice point with the shortest distance from the input data based on a preset quantization function; establishing an index corresponding to the input data and the lattice point, and mapping the input data to the lattice point; The method also includes: dividing the first lattice and the second lattice into a plurality of Voronoi units according to the generator matrix and the adjusted generator matrix; randomly generating a preset number of sample points in a preset unit hypercube; determining the lattice points mapped by the sample points in different lattices and the Voronoi units of the sample points in different lattices through the preset quantization function; counting the number of sample points in the Voronoi units in different lattices, respectively calculating the ratio of the number of sample points in each Voronoi unit to the preset number of sample points, determining the ratio as the volume of the corresponding lattice Voronoi unit, dividing the volumes of the Voronoi units of the two lattices, and obtaining the nesting ratio of the nested lattice system.

2. The method according to claim 1, characterized in that The steps before setting the Gaussian distribution function based on the nested lattice system include: Setting the theta series for the nested lattice system; Traversing the lattice points of the nested lattice system, and calculating theta-order values ​​of the nested lattice system; The parameters of the theta series are adjusted according to the theta series value until a first preset condition is met.

3. The method according to claim 2, characterized in that The setting of the Gaussian distribution function based on the nested lattice system comprises: When the first preset condition is met, determining the variance and the center of the Gaussian distribution; The lattice Gaussian distribution function is constructed according to the variance and the Gaussian distribution center.

4. The method according to claim 1, characterized in that: The calculating of a discrete Gaussian distribution group according to the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution comprises: Set up the initial data set; For each of the lattice points, dividing the single-point Gaussian distribution of the lattice point by the overall Gaussian distribution to obtain a discrete Gaussian distribution of the lattice point; The discrete Gaussian distribution of the lattice points is placed at corresponding positions of the initial data group to generate the discrete Gaussian distribution group.

5. The method according to claim 1, characterized in that: The method further comprises: In any lattice, for each of the sample points, calculate the square value of the distance from each of the sample points to the mapped lattice point; Accumulate the distance square value corresponding to each of the samples and divide it by the preset number to generate the second-order moment of the Voronoi unit; A variance of the Voronoi cell is generated based on the second order moment, the dimension of the basis vectors, and the volume of the Voronoi cell.

6. A simulation service platform data feature extraction device, characterized in that: The device comprises: A construction module, used to construct a nested lattice system; the lattice points of the nested lattice system have corresponding data; A function module, used for constructing a lattice Gaussian distribution function based on the nested lattice system; A first calculation module, configured to calculate a single-point Gaussian distribution for any of the lattice points based on the lattice Gaussian distribution function; A second calculation module, configured to add the single-point Gaussian distribution of each of the lattice points to obtain an overall Gaussian distribution of the nested lattice system; A third calculation module, used for calculating a discrete Gaussian distribution group according to the single-point Gaussian distribution of each lattice point and the overall Gaussian distribution; An extraction module, configured to extract key features of the data in the nested lattice system based on the discrete Gaussian distribution group, wherein the key features are used to cluster the data in the nested lattice system; The construction module includes: a basis vector submodule, which is used to obtain preset basis vectors and construct a generator matrix based on the preset basis vectors; a first lattice submodule, which is used to generate a first lattice according to the generator matrix; a second lattice submodule, which is used to adjust the generator matrix and generate a second lattice based on the adjusted generator matrix; the second lattice is a subset of the first lattice; and a nested lattice submodule, which is used to generate a nested lattice system according to the first lattice and the second lattice. The device further includes: an input module for acquiring input data; a quantization module for determining a lattice point with a minimum distance from the input data based on a preset quantization function; a mapping module for establishing an index corresponding to the input data and the lattice point, and mapping the input data to the lattice point; The device also includes: a division module, which is used to divide the first lattice and the second lattice into a plurality of Voronoi units according to the generation matrix and the adjusted generation matrix; a random module, which is used to randomly generate a preset number of sample points in a preset unit hypercube; a determination module, which is used to determine the lattice points mapped by the sample points in different lattices and the Voronoi units of the sample points in different lattices through a preset quantization function; and a statistical module, which is used to count the number of sample points in the Voronoi units in different lattices, respectively calculate the ratio of the number of sample points in each Voronoi unit to the preset number of sample points, determine the ratio as the volume of the corresponding lattice Voronoi unit, and divide the volumes of the Voronoi units of the two lattices to obtain the nesting ratio of the nested lattice system.

7. An electronic device, characterized in that: include: processor; as well as A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 5.

Images