Optimal optimization area identification method, device, equipment, medium and product

Abstract

The invention discloses an optimal optimization region identification method, device, equipment, medium and product, and relates to the field of data processing and machine learning, and the method comprises the steps: carrying out the preprocessing of a sample, carrying out the principal component analysis of the preprocessed sample, generating a plurality of two-dimensional spaces, respectively projecting each sample to each two-dimensional space, and generating a plurality of projection drawings; marking the sample projection points in the projection drawing to obtain optimal samples according to classification requirements; dividing a rectangular range formed by boundaries of sample projection points in each projection drawing into uniform grids, and counting the number of optimal samples and total samples in each grid; traversing all rectangular regions which can be formed by the grids in the projection drawing, and evaluating the score of each rectangular region by using a scoring function; and selecting the rectangular region with the highest evaluation score as the optimal optimization region, and outputting the boundary coordinates of the optimal optimization region, the proportion of the optimal samples in the total samples and the region area.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing and machine learning, in particular to an optimal optimization region identification method, device, equipment, medium and product. BACKGROUND

[0002] In the field of material data mining, the performance of a material is often determined by multiple component variables and process parameters, forming a high-dimensional nonlinear relationship. For example, in alloy component design, the performance optimization of a multi-component alloy system faces significant challenges. In order to alleviate the prediction risk brought about by complex nonlinear relationships, in material design, an optimization region can be located to control the material design and performance prediction risk. Within this optimization region, there are more samples, and the proportion of high-performance optimal samples is higher, and the proportion of low-performance poor samples is lower.

[0003] How to automatically identify the optimal optimization region in multiple projection maps is a challenge. Current optimization region identification mainly uses dimension reduction techniques such as principal component analysis (PCA) and linear discriminant analysis (LDA) for dimension reduction, and combines manual discrimination to find the optimal optimization region. For example, PCA projects high-dimensional data into low-dimensional space through orthogonal transformation, and LDA maximizes the class separability, but these methods have limitations when dealing with multiple projection maps: when the projection method (such as different dimension reduction algorithms or kernel functions) and the number of projection maps increase, manual discrimination needs to evaluate the distribution characteristics of each projection space one by one, resulting in low efficiency and being easily disturbed by subjective factors. In addition, there is a calculation bottleneck, and the complexity of exhaustive search is extremely high, which causes the algorithm to consume a lot of time when dealing with multiple projections or large amounts of data. SUMMARY

[0004] The purpose of the present application is to provide an optimal optimization region identification method, device, equipment, medium and product, which solves the problem of low efficiency and easy error of manual discrimination, realizes automatic identification of the optimal optimization region, and improves the identification accuracy and calculation efficiency.

[0005] To achieve the above purpose, the present application provides the following solutions: In a first aspect, the present application provides an optimal optimization region identification method, comprising: preprocessing the samples, generating a plurality of two-dimensional spaces from the preprocessed samples through principal component analysis, and projecting each sample into each two-dimensional space to generate a plurality of projection maps; labeling the optimal samples in the sample projection points in the projection maps according to the classification requirements; dividing the rectangular range composed of the sample projection points in each projection map into a uniform grid, and counting the number of optimal samples and total samples in each grid; traversing all rectangular areas that can be formed by the grids in the projection maps, and evaluating the score of each rectangular area using a scoring function; selecting the rectangular region with the highest evaluation score as the best optimization region, and outputting boundary coordinates of the best optimization region, a proportion of optimal class samples in total samples, and an area of the region.

[0006] Optionally, after the best optimization region is determined, the best optimization region identification method further comprises: drawing a rectangular region according to the boundary in the optimal projection map, and dividing the projection space into the best optimization region and a non-optimization region; wherein the optimal projection map is a projection map in which the best optimization region is located. combining the principal component analysis model and the standardized parameter, and converting the boundary of the best optimization region into a linear equation in the original feature space, so as to facilitate parameter control in actual application.

[0007] Optionally, the pre-processing of the samples comprises: standardizing the samples; dimensionally reducing the standardized samples by using the principal component analysis method, extracting the first n principal components, and combining the first n principal components two by two to form multiple two-dimensional spaces.

[0008] Optionally, the marking of the optimal class samples according to the classification requirement comprises: if the target performance is a discrete value, marking the specified optimal class samples; if the target performance is a continuous value, marking the optimal class samples according to a target performance range of the specified optimal class samples.

[0009] Optionally, the dividing of the rectangular range formed by the boundary of the sample projection points in each projection map into uniform grids comprises: for each projection map, dividing a rectangular range formed by four points at maximum and minimum values of x-axis and y-axis among all sample projection points in the projection map into uniform grids; wherein the number of grids in each projection map is the same.

[0010] Optionally, the specific calculation formula of the score function is: ; wherein the area A of the region is the number of grid units that constitute each rectangular region, is an optimal class proportion index, and p is a proportion of optimal class samples in total samples in the corresponding rectangular region.

[0011] In a second aspect, the present application provides a best optimization region identification device, comprising: a generating unit configured to pre-process samples, generate multiple two-dimensional spaces by principal component analysis on the pre-processed samples, and project each sample into each two-dimensional space to generate multiple projection maps. a marking unit configured to mark the sample projection points in the projection map as superior samples according to a classification requirement; a grid division unit configured to divide a rectangular range formed by the boundaries of the sample projection points in each projection map into uniform grids, and count the number of superior samples and total samples in each grid; an evaluation unit configured to traverse all rectangular regions formed by the grids in the projection map, and evaluate the score of each rectangular region using a scoring function; an output unit configured to select a rectangular region with the highest evaluation score as an optimal optimization region, and output the boundary coordinates of the optimal optimization region, the proportion of superior samples to total samples, and the area of the region.

[0012] In a third aspect, the present application provides a computer device, comprising a memory, a processor, a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the steps of the optimal optimization region identification method according to any one of the above.

[0013] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the optimal optimization region identification method according to any one of the above.

[0014] In a fifth aspect, the present application provides a computer program product, comprising a computer program, and the computer program is executed by a processor to implement the steps of the optimal optimization region identification method according to any one of the above.

[0015] According to the embodiments provided in the present application, the following technical effects are disclosed: The present application provides an optimal optimization region identification method, device, equipment, medium and product, which reduces the dimension of input data by using principal component analysis method, maps the high-dimensional feature space into multiple two-dimensional projections, reduces the data complexity while retaining the main feature information; through grid division and scoring function design, the algorithm complexity is controlled to be linear level, compared with traditional exhaustive search method, the calculation efficiency is improved by orders of magnitude when processing a large amount of data and multiple projections; at the same time, the method of the present application avoids the subjective factor interference caused by manual discrimination, and calculates based on objective scoring standard, which significantly improves the accuracy and automation degree of the optimal optimization region identification, and the advantage is particularly obvious in the big data scene, which can quickly respond to complex and changeable data environment, and provide timely and effective decision support for actual production and research. BRIEF DESCRIPTION OF DRAWINGS

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating an optimal region identification method provided in an embodiment of this application; Figure 2 A schematic diagram of the optimal optimization region with a superior category ratio index of 3 provided in an embodiment of this application; Figure 3 A schematic diagram of the optimal optimization region with a superior category ratio index of 1 provided in an embodiment of this application; Figure 4 This is a schematic diagram of the functional modules of an optimal region identification device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] In one exemplary embodiment, such as Figure 2 As shown, an optimal optimization region identification method is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, it includes the following steps 101 to 105. Wherein: Step 101: Preprocess the samples, generate multiple two-dimensional spaces by principal component analysis of the preprocessed samples, and project each sample onto each two-dimensional space to generate multiple projection maps.

[0021] In this embodiment, the samples used are obtained based on melting point performance datasets of multi-element alloys collected from databases and literature. These datasets include the chemical composition of the multi-element alloys. (i = 1, 2,..., N) and the corresponding melting point performance wherein the composition consists of m elements, the composition may be expressed as the mole percentage of each element is a vector of the composition, that is, satisfies = 1. In the embodiments of the present application, one sample represents the chemical composition of one alloy in the melting point performance data set of the multi-component alloy and the melting point performance thereof.

[0022] As an optional implementation, the pre-processing of the samples is performed to generate a plurality of two-dimensional spaces through principal component analysis of the pre-processed samples, and specifically includes: standardizing the samples; dimension reduction is performed on the standardized samples through the principal component analysis method, and then the first a principal components are extracted, and the first a principal components are combined two by two to form a plurality of two-dimensional spaces.

[0023] In the embodiments of the present application, the standardization of the samples refers to Z-score standardization of the composition of the alloy, and the specific formula is as follows: ; In the formula, is the characteristic mean, is the standard deviation.

[0024] In the embodiments of the present application, when the dimension reduction is performed through the principal component analysis method, the first 5 principal components are extracted, combined two by two to form 10 two-dimensional spaces, and the samples (one sample corresponds to one alloy composition) are projected into each two-dimensional space to obtain the coordinates of each sample in each two-dimensional space, and 10 projection maps are generated. In another embodiment, other number of principal components can be extracted, for example, 6, and finally 15 two-dimensional spaces can be formed, but it should be noted that since the number of generated two-dimensional spaces increases in steps with the number of extracted principal components, in order to avoid excessive calculation in the subsequent steps, the number of extracted principal components should not be too large.

[0025] In step 102, the sample projection points in the projection map are marked as optimal class samples according to the classification requirements; As an optional implementation, the sample projection points in the projection map are marked as optimal class samples according to the classification requirements, and specifically includes: If the target performance is a discrete value, the specified optimal class samples are marked; If the target performance is a continuous value, the target performance range of the specified optimal class samples is marked as the optimal class samples.

[0026] In the implementation of the embodiment, by marking the superior sample of discrete and continuous data, different types of task requirements can be adapted. In practical applications, whether it is a discrete judgment of whether the material is qualified or not, or the optimization of continuous performance indicators such as alloy melting point and strength, the optimization area can be accurately identified by reasonably setting the superior sample standard.

[0027] In the embodiment of the application, since the target performance is a continuous value, in the sample projection points in the projection map, the alloy samples with a melting point lower than 30℃ (accounting for 25% of the total samples) are marked as low-melting-point superior samples, and the remaining samples are non-superior samples.

[0028] Step 103, dividing the rectangular range composed of the boundaries of the sample projection points in each projection map into a uniform grid, and counting the number of superior samples and total samples in each grid; As an optional implementation, the dividing of the rectangular range composed of the boundaries of the sample projection points in each projection map into a uniform grid specifically includes: For each projection map, the rectangular range composed of the four points located at the maximum and minimum values of the x-axis and y-axis in all sample projection points of the projection map is divided into uniform grids; wherein the number of grids in each projection map is the same.

[0029] In the embodiment of the application, for each projection map, the rectangular range composed of the four points located at the maximum and minimum values of the x-axis and y-axis in all sample projection points of the projection map is uniformly divided into 20x20 grids, and the number of superior samples and the total number of samples in each grid are counted.

[0030] Step 104, traversing all the rectangular regions that can be formed by the grids in the projection map, and using a scoring function to evaluate the score of each rectangular region; In the embodiment of the application, the specific range of traversing all the rectangular regions is from 1x1 to 20x20 grids.

[0031] In the implementation of the embodiment, by designing the scoring function, the algorithm complexity is controlled to be linear, and compared with traditional exhaustive search and other methods, the calculation efficiency is improved by orders of magnitude when processing a large amount of data and multiple projections. At the same time, this method avoids the subjective interference caused by manual discrimination, and calculates based on objective scoring standards, which significantly improves the accuracy and automation degree of identifying the best optimization area.

[0032] As an optional implementation, the specific calculation formula of the scoring function is: ; Wherein, the area A of the region is the number of grid units that make up each rectangular region, and the area of each grid is the standardized projection space area; is a proportion index of the superior class, used to control the proportion between the superior class sample and the total sample in the importance of the score; p is the proportion of the superior class sample in the total sample in the corresponding rectangular region.

[0033] In this embodiment, by adjusting the proportion index of the superior class sample in the score function n , the weight of the superior class region area and the superior class proportion can be flexibly balanced. In the scene of pursuing high superior class sample density and extremely high performance accuracy requirement, the value of p can be increased to highlight the importance of the superior class sample proportion, and the small range region with high concentration of superior class samples is screened out; while in the scene of needing to ensure a certain sample quantity and expand the optimization selection space, the value of p is reduced to balance the area and proportion relationship, and a more inclusive optimization interval is obtained, so as to meet the diversified optimization target. n n

[0034] Step 105, selecting the rectangular region with the highest evaluation score as the best optimization zone, and outputting the boundary coordinates of the best optimization zone, the proportion of the superior class sample to the total sample and the area of the region. The projection map where the best optimization zone is located is the optimal projection map.

[0035] By implementing the above steps 101 to 105, the present application uses the principal component analysis method to perform dimension reduction processing on the input data, maps the high-dimensional feature space into multiple two-dimensional projections, reduces the data complexity while retaining the main feature information; through grid division and score function design, the algorithm complexity is controlled to be linear, compared with the traditional exhaustive search method, when processing a large amount of data and multiple projections, the calculation efficiency is improved by orders of magnitude; at the same time, the method of the present application avoids the subjective factor interference caused by manual discrimination, and calculates based on the objective scoring standard, which significantly improves the accuracy and automation degree of the best optimization zone identification, and the advantage is particularly obvious in the big data scene, which can quickly respond to the complex and changeable data environment, and provide timely and effective decision support for actual production and research.

[0036] As an optional embodiment, after the best optimization zone is determined, the best optimization zone identification method further comprises: Step 106, drawing a rectangular region according to the boundary in the optimal projection map, dividing the optimal projection map into the best optimization zone and the non-optimization zone; combining the principal component analysis (PCA) model and the standardization parameter, converting the boundary of the projection region into a linear equation in the original feature space to obtain the optimization zone boundary equation, which is convenient for parameter control in actual application.

[0037] ​​In the implementation, the mathematical control equation (optimization region boundary equation) of the optimal optimization region can be directly obtained, which can be combined with the principal component analysis (PCA) model and the standardized parameters to accurately convert the optimization region boundary in the low-dimensional projection space into a linear equation of the original feature space. For example, in the alloy composition optimization of the embodiment, the quantitative relationship and value range between different element contents can be intuitively presented, and researchers or engineers can clearly understand the influence mechanism of each feature variable on the optimization target without complex derivation process, which is convenient for direct application in actual production and material design, effectively guides parameter control and process adjustment, and greatly improves the practicability and reliability of the method.

[0038] In the optimal projection map, the optimal optimization region is marked with a red rectangle, as shown in Figure 2 The lower left corner coordinates are set as (x1, y1) , , and the upper right corner coordinates are set as (x2, y2) , . The relationship between the principal components and the original features is substituted into the principal component boundary equation (PC1 <PC1< , <PC2< ), to obtain the optimization region boundary equation. The relationship between the principal components and the original features is that the original features are linearly combined to obtain the principal components, PC1 is the first principal component, and PC2 is the second principal component.

[0039] Embodiment 1 The data of 150 low-melting-point alloys with different compositions reported in the literature were collected, and part of the data is shown in Table 1.

[0040] Table 1. Low-melting-point alloy data table

[0041] By using the optimal optimization region identification method provided in the application, when the optimal class proportion index is set to 3, the optimal optimization region and the optimal projection map are as shown in Figure 2 , wherein the total number of samples in the optimal optimization region is 34, the number of optimal class samples is 34, and the proportion of optimal class samples in the total samples is 100%. The optimal optimization region boundary equation is finally obtained as follows: ; ; This embodiment illustrates that the optimal optimization region identification method of the application can obtain the optimal optimization region and the boundary equation in multiple projection maps.

[0042] Embodiment 2 Example 2 is basically the same as Example 1, except that the optimal class proportion index is adjusted to 1. The final generated optimal optimization region and the optimal projection map are shown in Figure 3 , where the total number of samples in the optimal optimization region is 42, the number of optimal class samples is 34, and the proportion of optimal class samples in the total samples is 81%.

[0043] As can be seen from Example 2, by adjusting the value of the optimal class proportion index, flexible automatic positioning of the optimal optimization region can be achieved, and the size of the optimization region and the proportion of optimal class samples in the total samples can be balanced.

[0044] Based on the same inventive concept, the embodiments of the present application also provide an optimal optimization region identification device for implementing the optimal optimization region identification method described above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more optimal optimization region identification device embodiments provided below can refer to the limitations of the optimal optimization region identification method in the above text, which will not be repeated here.

[0045] In an exemplary embodiment, as shown in Figure 4 , an optimal optimization region identification device is provided, which includes: A generating unit 201 is configured to pre-process samples, generate a plurality of two-dimensional spaces by principal component analysis on the pre-processed samples, and project each sample into each two-dimensional space to generate a plurality of projection maps; A marking unit 202 is configured to mark optimal class samples according to classification requirements in the sample projection points in the projection maps; A grid division unit 203 is configured to divide a rectangular range composed of boundaries of the sample projection points in each projection map into uniform grids, and count the number of optimal class samples and total samples in each grid; An evaluation unit 204 is configured to traverse all rectangular regions that can be formed by the grids in the projection maps, and evaluate the score of each rectangular region using a scoring function; An output unit 205 is configured to select the rectangular region with the highest evaluation score as the optimal optimization region, and output the boundary coordinates of the optimal optimization region, the proportion of optimal class samples in the total samples, and the area of the region.

[0046] In the implementation, the input data is processed by dimension reduction through a principal component analysis method, the high-dimensional feature space is mapped into multiple two-dimensional projections, the data complexity is reduced while the main feature information is retained, the algorithm complexity is controlled to be linear, and compared with a traditional exhaustive search method, the calculation efficiency is improved by orders of magnitude when a large amount of data and multiple projections are processed. Meanwhile, the method avoids subjective interference caused by manual discrimination, performs calculation based on an objective scoring standard, significantly improves the accuracy and automation degree of the best optimization zone identification, and has obvious advantages in a big data scenario, can quickly respond to a complex and changeable data environment, and provides timely and effective decision support for actual production and research.

[0047] In an example embodiment, a computer device is provided, which can be a server or a terminal, and an internal structure diagram of the computer device can be as shown in Figure 5 The computer device includes a processor, a memory, an input / output interface (I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is configured to provide calculation and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for running of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is configured to store best optimization zone identification data. The input / output interface of the computer device is configured to exchange information between the processor and external devices. The communication interface of the computer device is configured to communicate with external terminals through network connection. The computer program is executed by the processor to implement a best optimization zone identification method.

[0048] Those skilled in the art can understand that Figure 5 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0049] In an example embodiment, a computer device is provided, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0050] In an example embodiment, a computer readable storage medium is provided, which stores a computer program. The computer program is executed by a processor to implement the steps in the above method embodiments.

[0051] In an example embodiment, a computer program product is provided, including a computer program which, when executed by a processor, implements the steps of the above-mentioned method embodiments.

[0052] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0053] It can be understood by those skilled in the art that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. Any reference to memory, database or other medium used in the embodiments provided by the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0054] The database involved in the embodiments provided by the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on blockchain, etc., without being limited thereto. The processor involved in the embodiments provided by the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0055] Any technical features in the above embodiments can be combined, and for the sake of brevity, not all possible combinations are described above, however, it should be understood that the application encompasses all possible combinations of the technical features unless such a combination is not technically possible.

[0056] The principles and implementation manners of the present application are described herein by using specific examples, and the above embodiments are only used to help understand the method of the present application and its core idea; meanwhile, according to the idea of the present application, the specific implementation manners and application scopes will be changed by those skilled in the art. In conclusion, the content of the present specification should not be understood as a limitation of the present application.

Claims

1. A method for identifying an optimal optimization zone, comprising: determining a first optimization zone; determining a second optimization zone; and determining a third optimization zone. The optimal optimization region identification method comprises: The sample is pretreated, and the pretreated sample is subjected to principal component analysis to generate a plurality of two-dimensional spaces, and each sample is projected into each two-dimensional space to generate a plurality of projection maps; The sample projection points in the projection map are marked as optimal class samples according to the classification requirements; The rectangular range formed by the boundaries of the sample projection points in each projection map is divided into a uniform grid, and the number of optimal class samples and total samples in each grid is counted; All rectangular regions that can be formed by the grid in the projection map are traversed, and the score of each rectangular region is evaluated using a scoring function; The rectangular region with the highest evaluation score is selected as the optimal optimization region, and the boundary coordinates of the optimal optimization region, the proportion of optimal class samples to total samples, and the area of the region are output.

2. The optimal optimization region identification method of claim 1, wherein, After the optimal optimization region is determined, the optimal optimization region identification method further comprises: A rectangular region is drawn according to the boundary in the optimal projection map to divide the projection space into an optimal optimization region and a non-optimization region; wherein the optimal projection map is the projection map in which the optimal optimization region is located; The boundary of the optimal optimization region is converted into a linear equation in the original feature space in combination with the principal component analysis model and the standardization parameter, facilitating parameter control in actual application.

3. The optimal optimization region identification method of claim 1, wherein, The sample is pretreated, and the pretreated sample is subjected to principal component analysis to generate a plurality of two-dimensional spaces, and each sample is projected into each two-dimensional space to generate a plurality of projection maps; The sample is standardized; The dimension of the sample after standardization is reduced by a principal component analysis method, and then the first α principal components are extracted, and the first α principal components are combined two by two to form a plurality of two-dimensional spaces.

4. The optimal optimization region identification method of claim 1, wherein, The sample projection points in the projection map are marked as optimal class samples according to the classification requirements; If the target performance is a discrete value, the specified optimal class samples are marked; If the target performance is a continuous value, the target performance range of the specified optimal class samples is marked as the optimal class samples.

5. The optimal optimization region identification method of claim 1, wherein, The rectangular range formed by the boundaries of the sample projection points in each projection map is divided into a uniform grid, and the number of optimal class samples and total samples in each grid is counted; For each projection map, the rectangular range formed by the four points at the maximum and minimum values of the x-axis and y-axis among all sample projection points of the projection map is divided into a uniform grid; wherein the number of grids in each projection map is the same.

6. The optimal optimization region identification method of claim 1, wherein, The specific calculation formula of the scoring function is: ； wherein the area A of the region is the number of grid cells that make up the respective rectangular region, is the proportion of the superior class, and p is the proportion of the superior class samples in the total samples in the corresponding rectangular region.

7. An optimal region identification device, characterized in that, The optimal optimization region identification device comprises: A generating unit is configured to pretreat the sample, generate a plurality of two-dimensional spaces by subjecting the pretreated sample to principal component analysis, and project each sample into each two-dimensional space to generate a plurality of projection maps; A marking unit is configured to mark the sample projection points in the projection map as optimal class samples according to the classification requirements; A grid division unit is configured to divide the rectangular range formed by the boundaries of the sample projection points in each projection map into a uniform grid, and count the number of optimal class samples and total samples in each grid; An evaluation unit is configured to traverse all rectangular regions that can be formed by the grid in the projection map, and evaluate the score of each rectangular region using a scoring function; An output unit is configured to select the rectangular region with the highest evaluation score as the optimal optimization region, and output the boundary coordinates of the optimal optimization region, the proportion of optimal class samples to total samples, and the area of the region.

8. A computer device comprising: Memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the optimal optimization region identification method of any one of claims 1-6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the optimal optimization region identification method of any one of claims 1-6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the optimal optimization region identification method of any one of claims 1-6.

Images