Method and device for determining cloud product solution and nonvolatile storage medium

Abstract

The invention discloses a method and device for determining a cloud product solution and a nonvolatile storage medium. The method comprises the steps that initial demand information is acquired, and the initial demand information comprises the business of the application cloud product and the resource configuration of the cloud product; under the condition that the request information is received, reference information is obtained, supplementary processing is conducted on the initial demand information through the reference information, target demand information is obtained, and the reference information comprises a plurality of historical schemes put into application and historical services corresponding to each historical scheme; each historical scheme comprises resource configuration of a plurality of target cloud products, and the request information is used for requesting to output a cloud product solution meeting the initial demand information; and determining a cloud product combination meeting the initial demand information in a plurality of public cloud products provided by the public cloud according to the target demand information, and outputting the cloud product combination as a cloud product solution.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular, to a method and device for determining a cloud product solution, and a non-volatile storage medium. Background Art

[0002] In the process of promoting cloud products, cloud service companies need to first collect accurate product requirements from the client, then design cloud product solutions that meet the product requirements based on the product requirements provided by the customer, and simultaneously provide a quotation for the cloud product solution. At present, in the actual demand matching process, the client may not be able to provide clear product requirements due to a low level of understanding of public cloud products and / or a low level of business mastery of the application of cloud products, resulting in the inability to accurately output cloud product solutions that meet product requirements; in addition, cloud service companies involve a large number of repeated calculations in the process of quotation calculation, which leads to problems such as long time consumption and low data processing efficiency.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present application provide a method and device for determining a cloud product solution, and a non-volatile storage medium, so as to at least solve the technical problem that a cloud product solution that meets the product requirements cannot be accurately output due to insufficient accuracy of the product requirements provided by the client.

[0005] According to one aspect of an embodiment of the present application, a method for determining a cloud product solution is provided, comprising: obtaining initial demand information, wherein the initial demand information includes: the business of applying the cloud product and the resource configuration of the cloud product; upon receiving request information, obtaining reference information, and using the reference information to supplement the initial demand information to obtain target demand information, wherein the reference information includes: multiple historical solutions that have been put into use, historical businesses corresponding to each historical solution, each historical solution containing resource configurations of multiple target cloud products, and the request information is used to request output of a cloud product solution that meets the initial demand information; determining a cloud product combination that meets the initial demand information from multiple public cloud products provided by a public cloud according to the target demand information, and outputting the cloud product combination as a cloud product solution, wherein the cloud product combination contains at least one public cloud product.

[0006] Optionally, the initial demand information is supplemented with reference information to obtain target demand information, including: determining the similarity between every two historical solutions, and generating a similarity matrix based on multiple similarities; generating multiple first eigenvectors based on the similarity matrix, and clustering the multiple first eigenvectors based on the distances between different first eigenvectors to obtain multiple first-category data sets, wherein each first eigenvector represents a historical solution, and each first-category data set contains multiple first eigenvectors; the initial demand information is supplemented with multiple first-category data sets to obtain a first supplementary result; the first eigenvectors are classified into multiple second-category data sets, wherein the number of second-category data sets is the same as the number of first-category data sets; the initial demand information is supplemented with multiple second-category data sets to obtain a second supplementary result; and the average of the first supplementary result and the second supplementary result is determined as the target demand information.

[0007] Optionally, multiple first eigenvectors are generated based on the similarity matrix, including: determining the cumulative results of multiple similarities corresponding to each historical scheme, and generating a first matrix with the multiple cumulative results as elements, wherein the first matrix is ​​a diagonal matrix; determining the difference between the first matrix and the similarity matrix as a second matrix, and normalizing the second matrix to obtain a normalized matrix, wherein the sum of multiple elements of each row of the normalized matrix is ​​a preset value; performing eigenvalue decomposition on the normalized matrix to obtain multiple eigenvalues, and multiple first eigenvectors corresponding to the multiple eigenvalues.

[0008] Optionally, cluster analysis is performed on multiple first eigenvectors, including: determining multiple target cluster centers among the multiple first eigenvectors, wherein each target cluster center is any first eigenvector, the number of target cluster centers is less than the number of first eigenvectors, and the number of target cluster centers is determined by iteratively clustering the multiple first eigenvectors; determining the distance between each first residual eigenvector and each target cluster center, and obtaining multiple distances corresponding to each first residual eigenvector, wherein the first residual eigenvector is the eigenvector other than the target cluster center; for each first residual eigenvector, determining the first target distance with the smallest value among the multiple distances corresponding to the first residual eigenvector, and classifying the first residual eigenvector and the target cluster center corresponding to the first target distance into a data set, wherein each data set contains a target cluster center and multiple first residual eigenvectors.

[0009] Optionally, the number of target cluster centers is determined by the following method: iteratively clustering multiple first feature vectors, wherein each time the iterative clustering is performed, the following loop process is performed: obtaining a random number, and determining a target feature vector as an initial cluster center from the multiple first feature vectors according to the random number, wherein the target feature vector is any first feature vector, and the random number is used to indicate the number of initial cluster centers; for each second residual feature vector, determining a first initial cluster center having the smallest distance to the second residual feature vector, and a second initial cluster center having a distance to the second residual feature vector that is only greater than the distance between the second residual feature vector and the first initial cluster center, wherein the second residual feature vector is the first feature vector other than the initial cluster center; determining The first distance between the second remaining eigenvector and each first eigenvector in the first data set to which the first initial cluster center belongs is determined, and a first average value of multiple first distances is determined, and a second distance between the second remaining eigenvector and each first eigenvector in the second data set to which the second initial cluster center belongs is determined, and a second average value of multiple second distances is determined; a first silhouette coefficient of each second remaining eigenvector is determined based on the first average value and the second average value, and the average value of the multiple first silhouette coefficients is determined as the second silhouette coefficient corresponding to the random number; when the number of iterative clustering reaches a preset number, a target silhouette coefficient with the largest value is determined among the multiple second silhouette coefficients corresponding to the multiple random numbers, and the target silhouette coefficient is determined as the number of target cluster centers.

[0010] Optionally, the initial demand information is supplemented with multiple data sets to obtain a first supplemented result, including: mapping the initial demand information into a second feature vector of the same dimension as the first feature vector, wherein the second feature vector contains a first target element whose mapping value is empty; determining the data set to which the target cluster center that is closest to the second feature vector among multiple target cluster centers belongs as the target data set; determining the target position of the first target element in the second feature vector, and determining the second target element indicated by each first feature vector in the target data set whose target position is determined, and determining the average value of multiple second target elements as the first supplemented element; filling the first supplemented element into the target position of the second feature vector to obtain the first supplemented result.

[0011] Optionally, the initial demand information is supplemented with multiple second-category data sets to obtain a second supplementary result, including: determining the posterior probability of the second eigenvector in each second-category data set, and determining the second-category data set corresponding to the posterior probability with the largest value as the third-category data set; determining the average value of all first eigenvectors in the third-category data set to obtain an average eigenvector; determining the element indicated by the target position in the average eigenvector as the second supplementary element, and supplementing the second supplementary element to the target position of the second eigenvector to obtain a second supplementary result.

[0012] Optionally, the first eigenvector is classified into a plurality of second-category data sets, comprising: randomly classifying the plurality of first eigenvectors into a plurality of initial data sets, wherein the number of the initial data sets is the same as the number of the first-category data sets; iteratively classifying the plurality of initial data sets, wherein each time the iterative classification is performed, the following loop process is performed: for each initial data set, an initial mean vector, an initial covariance matrix, and an initial weight coefficient of the initial data set are determined, wherein the initial mean vector is the average value of a plurality of first eigenvectors contained in the initial data set, and the elements in the initial covariance matrix include: a variance determined according to the initial data set, a covariance determined according to the initial data set, and an initial weight coefficient is used to indicate the number of first eigenvectors contained in the initial data set; and each first eigenvector is determined according to the initial mean vector, the initial covariance matrix, and the initial weight coefficient of each initial data set. The posterior probability of a feature vector in each initial data set, wherein the posterior probability is used to indicate the probability that the first feature vector is classified into the initial data set; determining a target posterior probability with the largest value among multiple posterior probabilities corresponding to the first feature vector, and reclassifying the first feature vector to the initial data set corresponding to the target posterior probability to obtain an updated result of the initial data set; when the difference between the first function value and the second function value is less than a preset value, determining that the iterative classification is completed, and determining multiple updated initial data sets output by the last iterative classification as the second type of data set, wherein the first function value is a function value determined according to the initial mean vector, initial covariance matrix and initial weight coefficient determined in the last iterative classification, and the second function value is a function value determined according to the initial mean vector, initial covariance matrix and initial weight coefficient generated in the next iterative classification.

[0013] Optionally, the method for determining a cloud product solution also includes: upon receiving a quotation request, determining each target cloud product included in the cloud product solution and determining a first quotation for each target cloud product; determining a second quotation for the cloud product solution based on the quotation of the target cloud product, and outputting the second quotation.

[0014] According to another aspect of an embodiment of the present application, a method for determining a cloud product solution is also provided, including: obtaining initial demand information, wherein the initial demand information includes: the business of applying the cloud product and the resource configuration of the cloud product; obtaining reference information, and using the reference information to supplement the initial demand information to obtain target demand information, wherein the reference information includes: multiple historical plans that have been put into use, historical business corresponding to each historical plan, and each historical plan contains resource configurations of multiple target cloud products; determining a cloud product combination corresponding to the initial demand information from multiple public cloud products provided by a public cloud according to the target demand information, and using the cloud product combination as the target cloud product solution, wherein the cloud product combination contains at least one public cloud product.

[0015] According to another aspect of an embodiment of the present application, a device for determining a cloud product solution is also provided, including: an acquisition module, used to obtain initial demand information, wherein the initial demand information includes: the business of applying the cloud product and the resource configuration of the cloud product; a supplementation module, used to obtain reference information when receiving request information, and use the reference information to supplement the initial demand information to obtain target demand information, wherein the reference information includes: multiple historical plans that have been put into use, historical business corresponding to each historical plan, each historical plan contains resource configurations of multiple target cloud products, and the request information is used to request the output of a cloud product solution that meets the initial demand information; a determination module, used to determine a cloud product combination that meets the initial demand information among multiple public cloud products provided by a public cloud according to the target demand information, and output the cloud product combination as a cloud product solution, wherein the cloud product combination contains at least one public cloud product.

[0016] According to another aspect of an embodiment of the present application, a non-volatile storage medium is provided, in which a computer program is stored, wherein the method for determining a cloud product solution is executed by running the computer program on a device where the non-volatile storage medium is located.

[0017] According to another aspect of an embodiment of the present application, an electronic device is also provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the above-mentioned method for determining a cloud product solution through the computer program.

[0018] According to another aspect of an embodiment of the present application, a computer program product is also provided, including computer instructions, which implement the steps of the above-mentioned method for determining a cloud product solution when the computer instructions are executed by a processor.

[0019] In an embodiment of the present application, initial demand information is obtained, wherein the initial demand information includes: the business of the application cloud product and the resource configuration of the cloud product; when the request information is received, reference information is obtained, and the initial demand information is supplemented with the reference information to obtain target demand information, wherein the reference information includes: multiple historical plans that have been put into use, historical business corresponding to each historical plan, and each historical plan contains resource configurations of multiple target cloud products, and the request information is used to request the output of a cloud product solution that meets the initial demand information; according to the target demand information, a cloud product combination that meets the initial demand information is determined from multiple public cloud products provided by the public cloud, and the cloud product combination is output as A cloud product solution, wherein the cloud product portfolio includes at least one public cloud product, and by supplementing the product requirements input by the client, product requirements containing more information (such as business type, resource configuration of cloud products) are obtained compared with the product requirements before supplementation, and the corresponding cloud product solution is determined based on the supplemented product requirements; wherein, by supplementing the product requirements input by the client, the purpose of improving the accuracy of the input information is achieved, thereby achieving the technical effect of improving the accuracy of the output cloud product solution, and further solving the technical problem of not being able to accurately output a cloud product solution that meets the product requirements due to insufficient accuracy of the product requirements provided by the client. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0021] Figure 1 is a hardware structure block diagram of a computer terminal for implementing a method for determining a cloud product solution according to an embodiment of the present application;

[0022] Figure 2 is a flowchart of steps of a method for determining a cloud product solution according to an embodiment of the present application;

[0023] Figure 3 is a flowchart of steps of another method for determining a cloud product solution according to an embodiment of the present application;

[0024] Figure 4 is a flow chart of supplementing initial demand information according to an embodiment of the present application;

[0025] Figure 5 is a structural diagram of a device for determining a cloud product solution according to an embodiment of the present application;

[0026] Figure 6This is a workflow diagram of a device for determining a cloud product solution according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] In order to better understand the embodiments of the present application, the technical terms involved in the embodiments of the present application are explained as follows:

[0030] Public cloud: A cloud computing service model in which cloud service providers provide computing resources, storage space, applications, and various services to the public or enterprises through the Internet.

[0031] Posterior probability: The posterior probability in the embodiment of the present application refers to the posterior probability under Gaussian distribution, which refers to the probability that the vector generated by the initial demand information mapping is classified into each second category data set.

[0032] In the related art, when supplementing the product requirements with vague business information or vague resource configuration information input by the client, the missing feature values ​​of the samples are directly filled in through the nearest neighbor samples, wherein all cloud product solutions in the case library are first clustered to obtain a set of feature values ​​with obvious similar features, and then the samples to be supplemented are further supplemented; this method, when the amount of data in the case library is small, leads to insufficient generalization ability of the clustering model when facing new data, and it is difficult to accurately divide new samples into appropriate clusters, resulting in inaccurate product requirements after supplementation, and further leading to inaccurate cloud product solutions output in the end. If clustering is not performed and the missing feature values ​​of the samples are directly filled in through the nearest neighbor samples, it is necessary to find the existing samples in the case library that are most similar to the input fuzzy requirements through the nearest neighbor algorithm, and use the features of the samples to fill in the missing features of the new samples; this method only considers the features of a single sample or several samples, and there is a problem of losing global information, resulting in inaccurate product requirements after supplementation, and it is impossible to output accurate cloud product solutions. In order to solve this problem, a relevant solution is provided in the embodiments of the present application, which is described in detail below.

[0033] According to an embodiment of the present application, a method embodiment for determining a cloud product solution is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0034] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal for implementing a method for determining a cloud product solution is shown. Figure 1 As shown, the computer terminal 10 may include one or more (102a, 102b, ..., 102n are used to illustrate) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0035] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10. As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0036] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the method for determining a cloud product solution in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the above-mentioned method for determining a cloud product solution is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0037] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0038] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 .

[0039] The present application embodiment provides a method for determining a cloud product solution that can be run in the above-mentioned operating environment. Figure 2 is a flowchart of the steps of the method for determining a cloud product solution provided in an embodiment of the present application, such as Figure 2 As shown, the method comprises the following steps:

[0040] Step S202, obtaining initial demand information, wherein the initial demand information includes: the business of applying the cloud product and the resource configuration of the cloud product.

[0041] The method provided in the embodiment of the present application is used to query cloud products that meet the product requirements in the cloud products provided by the public cloud for the product requirements input by the client, as a cloud product solution; or, the cloud products provided by the public cloud are combined, and the combination of multiple cloud products is used as a cloud product solution. Then in step S202, the demand information (i.e., initial demand information) input by the client needs to be obtained first, and the initial demand information is used to illustrate the configuration requirements of the cloud product and the application business requirements of the client side; the initial demand information usually includes: information used to illustrate the business of the application cloud product, for example, the business type (such as business field: finance, medical care, education); and the initial demand information also includes the resource configuration information of the cloud product, for example, the demand for computing resources, the demand for storage resources, and the demand for network resources. The initial demand information can also include: information such as company size and company nature, among which the information such as company size can be the number of employees, the number of salesmen, etc., which can be used to assist in determining the demand for the number of cloud products and the demand for the storage capacity of cloud products; the information such as company nature can be information such as educational institutions, medical institutions, etc. that represent the field of client application cloud products, which is used to assist in determining the business type of the application cloud product.

[0042] The method provided in the embodiment of the present application can also be executed using a trained neural network model. In step S202, the neural network model can be loaded into the memory. For example, the raw data of the neural network model can be loaded from the non-volatile memory into the volatile memory so that the processor can run the neural network model. The raw data of the neural network model refers to unprocessed data, which usually includes parameters and structural data of the neural network model. The structural data can be a calculation relationship based on the parameters, such as the forward propagation calculation relationship between the intermediate layers and between the neurons. Specifically, the structural data can include the code related to the structure of the neural network model, such as the code for performing related calculations between the intermediate layers and between the neurons.

[0043] In one embodiment, an area for loading a neural network model can be divided in the memory, which may include a structure data storage area and a parameter storage area. The structure data storage area is used to store structure-related codes, and the parameters referenced by it can point to the address of a specific parameter in the parameter storage area through a pointer. During the training process of the neural network model, it may be necessary to frequently update the parameters, and the parameter values ​​in the parameter storage area can be updated.

[0044] Step S204, upon receiving the request information, obtain reference information, and use the reference information to supplement the initial demand information to obtain the target demand information, wherein the reference information includes: multiple historical plans that have been put into use, historical businesses corresponding to each historical plan, and each historical plan contains resource configurations of multiple target cloud products. The request information is used to request the output of a cloud product solution that meets the initial demand information.

[0045] The method provided in the embodiment of the present application performs multiple supplementary processing on the fuzzy product demand information (i.e., initial demand information) input by the client, merges the product demands after being supplemented by different methods, and uses the result of the fusion as the supplemented product demand information. In this solution, not all initial demand information needs to be supplemented. For example, if the initial demand information includes additional information in addition to the business of the application cloud product and the resource configuration of the cloud product: such as the cloud host type and configuration information, object storage type and configuration information, elastic Internet Protocol Address (IP) number and shared bandwidth, World Wide Web (Web) protection, database and other value-added service information under the basic cloud product business, then the system / device that executes the method provided in the embodiment of the present application does not need to supplement the initial demand information after receiving the request information, but designs the public cloud product solution based on conventional design rules (i.e., product model, computing storage requirements, networking rules). When the initial demand information does not contain the above-mentioned additional information, in step S204, after receiving the request information, a supplementary processing process is executed to supplement the received initial demand information using different algorithms, wherein the request information may be a trigger signal for triggering the system / device that executes the method provided in the embodiment of the present application to output a cloud product solution that meets the initial demand information, and the cloud product solution that meets the initial demand information refers to a cloud product solution whose resource configuration is higher than or equal to the resource configuration information of the cloud product recorded in the initial demand information, and the cloud product solution can be applied to businesses including those recorded in the initial demand information. When supplementing the initial demand information, only the missing information of the initial demand information relative to the reference information is supplemented, that is, the initial demand information (i.e., the target demand information) after supplementation adds other information on the basis of retaining the initial demand information. In step S204, when the initial demand is supplemented, multiple historical solutions stored in the case library and multiple businesses (i.e., historical businesses) to which multiple historical methods are applied are obtained as reference information. The historical solution is a cloud product solution that has been output and fed back by the system / device that executes the method provided in the embodiment of the present application. Each historical solution records the resource configuration information of each cloud product (i.e., target cloud product) applied to it, for example, the configuration of computing resources, network resources, and storage resources of each cloud product (i.e., target cloud product) applied.

[0046] According to an optional embodiment of the present application, the initial demand information is supplemented with reference information to obtain target demand information, including: determining the similarity between every two historical solutions, and generating a similarity matrix based on multiple similarities; generating multiple first eigenvectors based on the similarity matrix, and clustering the multiple first eigenvectors based on the distances between different first eigenvectors to obtain multiple first-category data sets, wherein each first eigenvector represents a historical solution, and each first-category data set contains multiple first eigenvectors; supplementing the initial demand information with multiple first-category data sets to obtain a first supplementary result; classifying the first eigenvectors into multiple second-category data sets, wherein the number of second-category data sets is the same as the number of first-category data sets; supplementing the initial demand information with multiple second-category data sets to obtain a second supplementary result; and determining the average of the first supplementary result and the second supplementary result as the target demand information.

[0047] In this embodiment, reference information is screened from multiple historical solutions stored in the case library, and the reference information is used to supplement the initial demand information. For example, each historical solution in the case library is recorded as t, and the case library can be regarded as a sample set T containing multiple historical solutions, T = {t0, t1, ..., t n-1 When performing supplementary processing, case samples are first clustered according to the similarity between case samples. Specifically, a similarity algorithm (such as a Gaussian kernel function) is used to determine the similarity between every two case samples (i.e., historical plans), and a similarity matrix W is constructed based on multiple similarities, wherein the elements in the similarity matrix W are multiple similarity values, and each historical plan corresponds to multiple similarities (values). When clustering case samples, multiple eigenvectors (i.e., first eigenvectors) generated according to the similarity matrix W are clustered according to the similarity of the vectors to obtain a data set V containing multiple first eigenvectors. m (i.e. the first type of data set), and the data set V mThe similarity of any two first feature vectors in is greater than the preset similarity; wherein the similarity of the vectors can be determined by calculating the distance between the vectors, and the smaller the distance value of the two vectors, the higher the similarity. After the above-mentioned case sample clustering is completed, the first type of data set obtained by clustering is used as reference information to perform a supplementary processing on the initial demand information, and the initial demand information supplemented under this method (i.e., the first supplementary result) is obtained. The method provided in the embodiment of the present application needs to fuse the supplementary results obtained by different supplementary processing. In order to ensure information alignment and smooth fusion of the supplementary results, when another supplementary processing method is adopted, another data set (i.e., the second type of data set) with the same number as the first type of data set is generated based on multiple first feature vectors, and the second type of data set is used as reference information. Another supplementary processing algorithm is used to supplement the initial demand information using the second type of data set to obtain another supplementary processing result (i.e., the second supplementary result). During the fusion, the first supplementary result and the second supplementary result are averaged, and the average value is the result of the fusion, and is also the final supplementary result of the initial demand information (i.e., the target demand information).

[0048] Optionally, multiple first eigenvectors are generated based on the similarity matrix, including: determining the cumulative results of multiple similarities corresponding to each historical scheme, and generating a first matrix with the multiple cumulative results as elements, wherein the first matrix is ​​a diagonal matrix; determining the difference between the first matrix and the similarity matrix as a second matrix, and normalizing the second matrix to obtain a normalized matrix, wherein the sum of multiple elements of each row of the normalized matrix is ​​a preset value; performing eigenvalue decomposition on the normalized matrix to obtain multiple eigenvalues, and multiple first eigenvectors corresponding to the multiple eigenvalues.

[0049] In this embodiment, after constructing the similarity matrix W, it can also be normalized to obtain a normalized Laplace matrix L sym , further, calculate the Laplace matrix L sym The feature vectors {v0,v1,…,v n-1}, the Laplace matrix L sym After sorting the non-zero eigenvalues ​​in ascending order, select the eigenvectors {g0, g1, …, g m-1} as the first eigenvector. Wherein, the similarity matrix W is processed into a normalized Laplace matrix L symIn the process, the original Laplace matrix needs to be generated first. The specific method is as follows: the original Laplace matrix is ​​obtained by subtracting the similarity matrix and the degree matrix (i.e., the second matrix), wherein the degree matrix is ​​a diagonal matrix, and only the elements on the diagonal are not 0, and each element on its diagonal is the cumulative sum (i.e., the cumulative result) of all similarities (values) corresponding to a case sample; in this embodiment, after determining the multiple similarities corresponding to each case sample, the corresponding multiple similarities are accumulated to obtain a cumulative sum (i.e., the cumulative result); through the above method, each case sample (i.e., the historical scheme) has its corresponding cumulative result; further, a diagonal matrix (i.e., the first matrix) with multiple cumulative results as matrix elements is generated. Next, the similarity matrix W is subtracted from the first matrix constructed above, and the difference obtained is used as the original Laplace matrix (i.e., the second matrix). Next, the Laplace matrix (i.e., the second matrix) is normalized to obtain the normalized Laplace matrix L sym ; The normalized Laplacian matrix L sym The sum of each element in each row of is a preset value (for example, all 1). sym (i.e., normalized matrix) can be subjected to eigendecomposition to obtain multiple eigenvalues ​​and multiple (first) eigenvectors corresponding to the multiple eigenvalues. If each first eigenvector is denoted as g, then the characteristic matrix V m =[g0 g1…g m-1 ], m represents the number of (first) eigenvectors obtained by the eigendecomposition. In addition, after the eigendecomposition, the normalized Laplacian matrix L can be sym The non-zero eigenvalues ​​in the (i.e., normalized matrix) are sorted in ascending order, and the eigenvectors corresponding to the first n eigenvalues ​​are selected as the first-category eigenvectors.

[0050] According to another optional embodiment of the present application, a cluster analysis is performed on multiple first eigenvectors, including: determining multiple target cluster centers among the multiple first eigenvectors, wherein each target cluster center is any first eigenvector, the number of target cluster centers is less than the number of first eigenvectors, and the number of target cluster centers is determined by iteratively clustering the multiple first eigenvectors; determining the distance between each first residual eigenvector and each target cluster center, and obtaining multiple distances corresponding to each first residual eigenvector, wherein the first residual eigenvector is the eigenvector other than the target cluster center; for each first residual eigenvector, determining the first target distance with the smallest value among the multiple distances corresponding to the first residual eigenvector, and classifying the first residual eigenvector and the target cluster center corresponding to the first target distance into a data set, wherein each data set contains a target cluster center and multiple first residual eigenvectors.

[0051] After the similarity matrix W is subjected to eigendecomposition and a plurality of first eigenvectors are obtained, cluster analysis needs to be performed on them to obtain reference information for supplementary processing of the initial demand information, wherein the method provided in this embodiment can be referred to when clustering the first eigenvectors. In this embodiment, when clustering the plurality of first eigenvectors, first, a plurality of first eigenvectors (i.e., target cluster centers) serving as the central data of the first category of data sets are determined by iterative clustering, wherein the number of first eigenvectors determined as target cluster centers is determined by the iterative clustering in the previous step, and therefore, when clustering different data sets, the number of target cluster centers may be different. After the number of target cluster centers is determined as described above, the number of the first category of data sets is also determined. The number of the first category of data sets is the same as the number of target cluster centers. Therefore, when performing cluster analysis, each target cluster center is used as the center of the (first category) data set, and multiple first residual eigenvectors that are close to the target cluster center are determined from other first eigenvectors (i.e., first residual eigenvectors) that are not determined as target cluster centers. Each target cluster center and multiple first residual eigenvectors that are close to each target cluster center are combined into a (first category) data set. When determining multiple first residual eigenvectors that are close to the target cluster center, the Euclidean distance can be used. For each first residual eigenvector, its Euclidean distance with each target cluster center is determined respectively, and multiple Euclidean distances corresponding to a (first) residual eigenvector can be obtained. The Euclidean distance with the smallest value (i.e., the first target distance) and the target cluster center whose distance from this residual eigenvector is the above-mentioned target distance are classified into a (first category) data set; in this embodiment, the target cluster center is a fixed (first) eigenvector used to cluster other residual eigenvectors, and each first category data set contains only one target cluster center.

[0052] Optionally, the number of target cluster centers is determined by the following method: iteratively clustering multiple first feature vectors, wherein each time the iterative clustering is performed, the following loop process is performed: obtaining a random number, and determining a target feature vector as an initial cluster center from the multiple first feature vectors according to the random number, wherein the target feature vector is any first feature vector, and the random number is used to indicate the number of initial cluster centers; for each second residual feature vector, determining a first initial cluster center having the smallest distance to the second residual feature vector, and a second initial cluster center having a distance to the second residual feature vector that is only greater than the distance between the second residual feature vector and the first initial cluster center, wherein the second residual feature vector is the first feature vector other than the initial cluster center; determining The first distance between the second remaining eigenvector and each first eigenvector in the first data set to which the first initial cluster center belongs is determined, and a first average value of multiple first distances is determined, and a second distance between the second remaining eigenvector and each first eigenvector in the second data set to which the second initial cluster center belongs is determined, and a second average value of multiple second distances is determined; a first silhouette coefficient of each second remaining eigenvector is determined based on the first average value and the second average value, and the average value of the multiple first silhouette coefficients is determined as the second silhouette coefficient corresponding to the random number; when the number of iterative clustering reaches a preset number, a target silhouette coefficient with the largest value is determined among the multiple second silhouette coefficients corresponding to the multiple random numbers, and the target silhouette coefficient is determined as the number of target cluster centers.

[0053] As mentioned in the previous embodiment, the number of target cluster centers is determined by iterative clustering, wherein the iterative clustering process is as shown in this embodiment, and the initial cluster center is updated in each iterative clustering process, so as to determine multiple target cluster centers after completing the last iterative clustering; wherein the number of initial cluster centers and the first eigenvectors as the initial cluster centers are randomly determined, for example, the multiple first eigenvectors obtained by eigendecomposition correspond to the characteristic matrix V m =[g0 g1…g m-1 ], in this embodiment, the row data is recorded as {r0, r1, ..., r n-1}, randomly select j new samples as the initial cluster centers [q0,q1,…,q j-1 ], in the first iteration of clustering, the value of j is equal to the generated random number. Next, use multiple initial cluster centers to classify the remaining first eigenvectors (i.e., second residual eigenvectors) that are not the initial cluster centers; for example, calculate the first eigenvectors that are not the initial cluster centers according to each second residual eigenvector and each initial cluster center [q0,q1,…,q j-1] of the Euclidean distance, each second remaining eigenvector is assigned to the initial cluster center closest to it, forming a data set corresponding to the initial cluster center; according to the above method, multiple data sets corresponding to multiple initial cluster centers are obtained. Further, after the last iterative clustering is completed, multiple data sets corresponding to multiple initial cluster centers are output, and in the next iterative clustering, the initial cluster centers used for clustering are updated, including: updating the initial cluster centers and updating the number of initial cluster centers, and the number of initial cluster centers is updated using the silhouette coefficient method as follows: first, determine the distance between each eigenvector that is not used as the initial cluster center (i.e., the second remaining eigenvector) and each initial cluster center to determine the initial cluster center closest to it (i.e., the first initial cluster center), and determine the initial center cluster with the second closest distance to it (i.e., the second initial cluster center); the above-mentioned "second closest distance" means that when the distances between each second remaining eigenvector and each initial cluster center are arranged in order from large to small, the sorting position is the second distance value. Next, for each eigenvector that is not the center of the initial cluster (i.e., the second remaining eigenvector), the silhouette coefficient of the second remaining eigenvector is determined based on the data set (i.e., the first data set) to which the initial cluster center that is closest to it (i.e., the first initial cluster center) belongs, and the data set (i.e., the second data set) to which the initial cluster center that is second closest to it (i.e., the second initial cluster center) belongs; specifically, the distance between each eigenvector that is not the center of the initial cluster and each eigenvector in the first data set corresponding to it is calculated, and the multiple distances obtained are averaged to obtain an average value (i.e., the first average value), and the distance between each eigenvector that is not the center of the initial cluster and each eigenvector in the second data set corresponding to it is calculated at the same time, and the multiple distances obtained are averaged to obtain another average value (i.e., the first average value). That is, the second average value). The above-mentioned first average value and second average value are the silhouette coefficients (that is, the first silhouette coefficients) of the feature vector that is not used as the initial cluster center. Since there are multiple feature vectors that are not used as the initial cluster centers, multiple first silhouette coefficients corresponding to multiple feature vectors that are not used as the initial cluster centers can be obtained in the above manner. The average value of these multiple first silhouette coefficients is the (second) silhouette coefficient corresponding to the number of initial cluster centers (that is, the random number) of this round of iterative clustering. Before reaching the preset number of iterations (that is, the preset number of times), each clustering iteration updates the number of cluster centers through the above-mentioned scheme. The initial cluster center is updated by the following method: the average value of all feature vectors in the data set to which each initial cluster center belongs is calculated, and the average value of multiple feature vectors corresponding to the multiple data sets is obtained as the new initial cluster center [q0 * ,q1 * ,…,q j-1 *]. Each iteration of clustering updates the initial cluster center of this iteration in this way. In the process of iterative clustering, the above process of updating the initial cluster center and updating the number of initial cluster centers is repeated until the number of iterations reaches a preset number, and the second silhouette coefficient with the largest value (i.e., the target silhouette coefficient) is determined from the multiple second silhouette coefficients output from multiple iterations, and is used as the number of target cluster centers.

[0054] According to some optional embodiments of the present application, the initial demand information is supplemented with multiple data sets to obtain a first supplemented result, including: mapping the initial demand information into a second feature vector of the same dimension as the first feature vector, wherein the second feature vector contains a first target element whose mapping value is empty; determining the data set to which the target cluster center closest to the second feature vector among multiple target cluster centers belongs as the target data set; determining the target position of the first target element in the second feature vector, and determining the second target element indicated by each first feature vector in the target data set whose target position is determined, and determining the average value of the multiple second target elements as the first supplemented element; filling the first supplemented element into the target position of the second feature vector to obtain the first supplemented result.

[0055] In some embodiments, when the first type of data set obtained by clustering is used as reference information to supplement the initial demand information, the initial demand information needs to be mapped to a feature vector (i.e., the second feature vector) with the same dimension as each feature vector (i.e., the first feature vector) in the first type of data set. For example, if the first feature vector is m-dimensional, the dimension of the feature vector (i.e., the first feature vector) obtained by mapping the initial demand information is also m; however, the initial demand information needs to be supplemented precisely because it lacks some information compared to the reference information. Therefore, when the initial demand information is mapped to a (second) feature vector with the same dimension as the feature vector (i.e., the first feature vector) representing the reference information, there must be an element (i.e., the first target element) with an empty mapping value in the (second) feature vector corresponding to the initial demand information, that is, the structure of the first feature vector and the second feature vector is the same (i.e., the number of elements is the same, and the elements in the same position represent the same meaning), but the second feature vector lacks meaningful elements relative to the first feature vector (i.e., elements whose mapping values ​​are not empty). In this embodiment, only the elements whose mapping values ​​are empty are supplemented, and the elements whose value results in the second eigenvector are not empty (i.e., the elements of recorded information) are retained. Therefore, the supplemented results obtained by the method provided in the embodiment of the present application include both the information added during the supplementary processing and the information originally recorded in the initial requirement information. After the initial demand information has been mapped to a feature vector, the feature vector representing the initial demand information (i.e., the second feature vector) and multiple feature vectors representing the reference information (i.e., the first feature vector) can be further clustered according to the distance. Specifically, multiple target cluster centers are randomly determined in the feature vector representing the reference information (i.e., the first feature vector), and the number of target cluster centers is determined by the iterative clustering method in the previous embodiment. Clustering is performed according to the distance between other feature vectors that are not the target cluster center (i.e., the first remaining feature vector) and the target cluster center, and the data set corresponding to each target cluster center can be obtained; for example, for a first remaining feature vector, its distance with each target cluster center is determined respectively, and the distance value with the smallest value among the obtained multiple distance values ​​is determined, and the target cluster center corresponding to the first remaining feature vector and the above-mentioned smallest distance value is obtained; through the above method, each first remaining feature vector is divided into the data set corresponding to the target cluster center with the smallest distance therefrom, and the data set with the same number of target cluster centers can be obtained. Next, all feature vectors contained in the set (i.e., the target data set) to which the target cluster center with the smallest distance to the feature vector representing the initial demand information (i.e., the second feature vector) belongs is used as information for supplementing the initial demand information.When performing supplementary processing, determine the element with empty mapping value (i.e., the first target element) in the feature vector (i.e., the second feature vector) representing the initial demand information, determine the position of the first target element in the second feature vector as the target position, determine the element at the same position in each first feature vector contained in the target data set, and obtain multiple elements (i.e., the second target elements) at the above target position in each first feature vector; use the average value of the multiple second target elements as the supplementary element at the target position of the first feature vector (i.e., the first supplementary element), fill it into the target position of the second feature vector, and complete the supplementation of the target position; after all the target positions with empty mapping values ​​in the second feature vector are supplemented, the initial demand information supplemented by the clustering method (i.e., the first supplementary result) can be obtained. For example, the feature vector t representing the initial demand information. n The element with an empty mapping value (i.e., the first target element) in (i.e., the second eigenvector) is located in the Mth row and Nth column of the second eigenvector, then the target position is (M, N); the element at position (M, N) in each first eigenvector is the second target element; the average value of these multiple second target elements will be filled into the second eigenvector t n At the target position, we get the completed sample t n1 (i.e. the first supplementary result).

[0056] It should be noted that the first residual feature vector and the second residual feature vector in the embodiment of the present application are essentially first feature vectors. They are classified into the first residual feature vector and the second residual feature vector when the first feature vector is screened using different screening rules (such as whether it is used as the initial cluster center, whether it is used as the target cluster center, etc.).

[0057] According to an optional embodiment of the present application, the initial demand information is supplemented with multiple second-category data sets to obtain a second supplemented result, including: determining the posterior probability of the second eigenvector under each second-category data set, and determining the second-category data set corresponding to the posterior probability with the largest value as the third-category data set; determining the average value of all first eigenvectors in the third-category data set to obtain an average eigenvector; determining the element indicated by the target position in the average eigenvector as the second supplemented element, and supplementing the second supplemented element to the target position of the second eigenvector to obtain a second supplemented result.

[0058] As mentioned in the above embodiments, in order to improve the accuracy of the supplementation results, the method provided in the embodiments of the present application simultaneously supplements the initial demand information by different methods. In addition to the clustering-based method in the above embodiments to supplement the initial demand information, the initial demand information is also supplemented based on probability classification. In this embodiment, when the initial demand information is supplemented based on probability classification, first, it is determined how many (second-category) data sets are generated by classifying the first feature vector based on probability classification. In this embodiment, the number of data sets obtained based on clustering (i.e., the first-category data sets) is unified with the number of data sets obtained based on probability classification (i.e., the second-category data sets), that is, the number of the second-category data sets is determined based on the number of the first-category data sets. If the number of the first-category data sets is j, the number of the second-category data sets is also j. Next, the posterior probability of the vector (i.e., the second eigenvector) generated by the initial demand information mapping is determined under each second-category data set, wherein the posterior probability of the second eigenvector under each second-category data set is the probability that the second eigenvector is classified into this second-category data set. Therefore, the eigenvector representing the initial demand information (i.e., the second eigenvector) is classified into the second-category data set (i.e., the third-category data set) corresponding to the posterior probability with the largest value among its corresponding multiple posterior probabilities, and the first eigenvector contained in the second-category data set (i.e., the third-category data set) is determined as the vector used to supplement the initial demand information under the probabilistic classification method. When the first eigenvector contained in the third category of data sets is used to supplement the initial demand information, all the first eigenvectors contained in the third category of data sets are averaged to obtain an average eigenvector; since the method provided in the embodiment of the present application only supplements the missing information, no matter which method is used for supplementation, it is necessary to determine the position (i.e., target position) of the element with an empty mapping value (i.e., the first target element) in the second eigenvector in the vector representing the initial demand information (i.e., the second eigenvector) to supplement the target position. In this embodiment, the element at the target position in the average eigenvector is used as the element (i.e., the second supplementary element) to supplement the initial demand information under the probability classification method, and it is filled into the target position of the vector representing the initial demand information, so that the supplementation result (i.e., the second supplementary result) under the probability classification method can be obtained.

[0059] Optionally, the first eigenvector is classified into a plurality of second-category data sets, comprising: randomly classifying the plurality of first eigenvectors into a plurality of initial data sets, wherein the number of the initial data sets is the same as the number of the first-category data sets; iteratively classifying the plurality of initial data sets, wherein each time the iterative classification is performed, the following loop process is performed: for each initial data set, an initial mean vector, an initial covariance matrix, and an initial weight coefficient of the initial data set are determined, wherein the initial mean vector is the average value of a plurality of first eigenvectors contained in the initial data set, and the elements in the initial covariance matrix include: a variance determined according to the initial data set, a covariance determined according to the initial data set, and an initial weight coefficient is used to indicate the number of first eigenvectors contained in the initial data set; and each first eigenvector is determined according to the initial mean vector, the initial covariance matrix, and the initial weight coefficient of each initial data set. The posterior probability of a feature vector in each initial data set, wherein the posterior probability is used to indicate the probability that the first feature vector is classified into the initial data set; determining a target posterior probability with the largest value among multiple posterior probabilities corresponding to the first feature vector, and reclassifying the first feature vector to the initial data set corresponding to the target posterior probability to obtain an updated result of the initial data set; when the difference between the first function value and the second function value is less than a preset value, determining that the iterative classification is completed, and determining multiple updated initial data sets output by the last iterative classification as the second type of data set, wherein the first function value is a function value determined according to the initial mean vector, initial covariance matrix and initial weight coefficient determined in the last iterative classification, and the second function value is a function value determined according to the initial mean vector, initial covariance matrix and initial weight coefficient generated in the next iterative classification.

[0060] As mentioned in the previous embodiment, in the process of supplementing the initial demand information based on the probability classification method, multiple second-class sets need to be generated. In the previous embodiment, it has been explained that multiple first feature vectors are randomly classified into multiple second-class data sets, and the number of second-class data sets is always consistent with the number of first-class data sets; Next, it is described in detail how to classify the first feature vectors into data sets (second-class data sets) for the next step of supplementing the initial demand information. In this embodiment, after the first random generation of multiple second-class data sets, these second-class data sets are regarded as initial data sets and iterated, wherein, in the first iteration, the above-mentioned randomly generated multiple second-class data sets are taken as iteration objects (i.e., initial data sets), and the objects iterated in each iteration thereafter are called initial data sets. In each iteration, the posterior probability of each first feature vector under each initial set is determined, so as to reclassify the first feature vector according to the posterior probability, and realize the update of the initial data set during the iteration process. The posterior probability is determined based on the Gaussian distribution. Therefore, before determining the posterior probability of each first feature vector under each initial data set, the relevant parameters of the Gaussian distribution are first determined. In this embodiment, the relevant parameters of the Gaussian distribution need to be updated in each iteration. The relevant parameters of the Gaussian distribution include: initial mean vector, initial covariance matrix and initial weight coefficient, wherein each initial data set generates a corresponding initial mean vector, initial covariance matrix and initial weight coefficient; each initial mean vector is the average value of multiple first eigenvectors contained in its corresponding initial data set; the elements in each initial covariance matrix are multiple covariances, which are determined by the following method: calculating the covariance of each first eigenvector and multiple first eigenvectors contained in each initial data set; the value of each initial weight coefficient is the same as the number of first eigenvectors contained in its corresponding initial data set. Next, the posterior probability of each first eigenvector is determined according to the Bayesian formula: Determine the first eigenvector t n The posterior probability P(t n ), where n represents the number of first eigenvectors (i.e., the number of case samples), d represents the dimension of the first eigenvector, Σ represents the covariance matrix, μ i represents the mean vector, Σ -1 is the inverse matrix of the covariance matrix Σ, (t n -μ i ) T It means (t n -μ i). After obtaining the posterior probability of a first eigenvector under multiple initial data sets through the above method, reclassify the first eigenvector into the initial data set corresponding to the posterior probability with the largest value (i.e., the target posterior probability); reclassify each first eigenvector by the above method of calculating the posterior probability, and after all the first eigenvectors are classified into the initial data set corresponding to the posterior probability with the largest value, an iteration is completed, and the reclassified multiple data sets are output as the results of this initial data set iteration. The above reclassification process is executed cyclically, and the likelihood function is calculated after each reclassification; in this embodiment, the change value of the likelihood function is used to determine whether to stop the loop; specifically, the likelihood function value calculated after the previous iteration (i.e., the first function value) is subtracted from the likelihood function value calculated after the next iteration (i.e., the second function value), and when the difference is less than the preset value (for example, 10 -4 ), the loop stops, and the multiple data sets output after the last iteration of classification are used as the data set (the second type of data set) for the next step of supplementary processing of the initial demand information. Likelihood function Where K represents the first eigenvector t n The dimension of N represents the total number of first eigenvectors, Σ -1 is the inverse matrix of the covariance matrix Σ, (t n -μ i ) T It means (t n -μ i ), μ i Represents the mean vector. For example, the process of classifying the initial demand information based on probability classification is as follows: The first eigenvector is t n , in the probability classification based on t n When performing supplementary processing, in sample t n Under the available features of the data and the dimensions of the available features, determine the initial mean vector μ for each of the j categories (i.e., j sets of second-category data) i (i∈[0,j-1]), determine the initialization covariance matrix ∑ i (i∈[0,j-1]), determine the initial weight coefficient π i Table t of samples in cluster i n Calculate the number of each sample data {t0,t1,…,t n-1}Posterior probability α under Gaussian distribution of j categories id (i∈[0,n-1],d∈[0,j-1]). Divide each sample data into the category with the highest posterior probability. Update the mean vector μ i (i∈[0,j-1]), update the covariance matrix ∑i (i∈[0,j-1]), update weight coefficient π i Repeat the above probability classification process until the change in the log-likelihood function is less than the preset value (10 -4 ). Calculate the posterior probability of the sample to be completed (i.e., the feature vector representing the initial demand information) under the j-classification Gaussian distribution, and classify it into the category with the highest probability; use the mean of the original data of the sample under the category where the vector representing the initial demand information is located to complete the missing feature data, and obtain the completed new sample t n2 .

[0061] If the initial demand information (i.e., the first supplementary result) after supplementation based on iterative clustering is t n1 , the initial demand information after supplementation based on probability classification (i.e., the second supplementation result) is t n2 , then t n1 and t n2 In the vector t corresponding to the initial demand information n At the feature position where data is missing (i.e., the target position), the average value is calculated and then filled in to obtain the final new sample (i.e., the target required information) t n3 .

[0062] Step S206, determining a cloud product combination that meets the initial demand information from a plurality of public cloud products provided by the public cloud according to the target demand information, and outputting the cloud product combination as a cloud product solution, wherein the cloud product combination includes at least one public cloud product.

[0063] After completing the supplementary processing of the initial demand information in step S204, in step S206, the corresponding cloud product solution is output based on the richer and more accurate product demand information (i.e., the target demand information); the cloud product solution in the present application is generated based on the public cloud scenario, and the method device / system provided in the embodiment of the present application selects a cloud product or a combination of multiple cloud products in the public cloud whose resource configuration and application services are consistent with the target demand information according to the computing storage requirements, networking rules, application services and other information of the cloud product recorded in the target demand information, and outputs the above-mentioned cloud product or a combination of multiple cloud products that are consistent with the target demand information as a cloud product solution; wherein, the cloud products provided by the public cloud include: cloud hosts, cloud hard disks, cloud desktops, object storage, etc., as well as value-added services such as databases.

[0064] According to some other optional embodiments of the present application, the method for determining a cloud product solution also includes: upon receiving a quotation request, determining each target cloud product included in the cloud product solution, and determining a first quotation for each target cloud product; determining a second quotation for the cloud product solution based on the quotation of the target cloud product, and outputting the second quotation.

[0065] The method provided in the embodiment of the present application can also output a quotation (i.e., a second quotation) for the cloud product solution after outputting a cloud product solution that meets the initial demand information (i.e., a solution that meets the target demand information, but when the initial demand information needs to be supplemented, the product design solution cannot be directly generated based on the initial demand information). Specifically, in this embodiment, if a request information (i.e., a quotation request) input by a client for requesting the price of a solution for outputting a cloud product is received, the quotation (i.e., the second quotation) for the cloud product solution is determined based on the individual quotation (i.e., the first quotation) of each cloud product included in the cloud product solution; specifically, the public cloud products (i.e., the target cloud products) included in the cloud product solution are determined, the individual quotation (i.e., the first quotation) of each cloud product is determined, and the cumulative sum of the multiple first quotations corresponding to the multiple target cloud products is output as the quotation (i.e., the second quotation) for the cloud product solution; or, the cumulative sum of the multiple first quotations corresponding to the multiple target cloud products is added with a preset value-added price, and the cumulative sum of the multiple first quotations and the value-added price is output as the quotation (i.e., the second quotation) for the cloud product solution.

[0066] Through the above steps, when the initial demand information input by the client is missing information relative to the reference information, different supplementary processing methods can be used to supplement the initial demand information, and the initial demand information supplemented by different supplementary processing methods can be merged to obtain the final result of supplementing the initial demand information, thereby increasing the amount of information in the initial demand information; the corresponding cloud product solution is output based on the supplemented initial information, thereby improving the accuracy of the output cloud product solution.

[0067] Figure 3 is a flowchart of another method for determining a cloud product solution according to an embodiment of the present application. Figure 3 As shown, the method comprises the following steps:

[0068] Step S302, obtaining initial demand information, wherein the initial demand information includes: the business of applying the cloud product and the resource configuration of the cloud product.

[0069] In step S302, the system / device executing the method provided by the embodiment of the present application obtains initial demand information input by the user from the interactive interface or the client, wherein the initial demand information is the user's demand for the cloud product, including the business that applies the cloud product and the resource configuration of the cloud product; such information on the business that applies the cloud product includes information such as the business type (such as business areas: finance, medical care, education); such information on the resource configuration of the cloud product includes information such as the demand for computing resources, the demand for storage resources, the demand for network resources, etc.

[0070] Step S304, obtain reference information, and use the reference information to supplement the initial demand information to obtain target demand information, wherein the reference information includes: multiple historical plans that have been put into use, historical businesses corresponding to each historical plan, and resource configurations of multiple target cloud products contained in each historical plan.

[0071] If the information recorded in the acquired initial demand information covers accurate business information, such as the cloud host model, object storage requirements, elastic IP, shared bandwidth, and database, it is determined that there is no need to supplement the initial demand information, and the corresponding product solution is directly generated based on the initial demand information. If the initial demand information only records the business type, the nature and scale of the object (such as an enterprise or company) that applies the cloud product, and the numerical range of the computing resources and the numerical range of the network resources, it is determined that the initial demand information needs to be supplemented. In this embodiment, the steps shown in step S304 are used to supplement the initial demand information. Figure 4 It is a flow chart for supplementing the initial demand information, such as Figure 4 As shown, in this embodiment, two different methods are used to supplement the initial demand information twice based on the same reference information, wherein the reference information includes: a case sample set consisting of multiple case samples, each case sample is a cloud product solution (i.e., historical solution) that has been put into use before the current moment, wherein each historical solution contains the resource configuration information of the cloud product (i.e., target cloud product) contained in this historical solution; the reference information also includes the business (i.e., historical business) corresponding to the cloud product solution (i.e., historical solution) that has been put into use before. Figure 4As shown, in a supplementary processing scheme (i.e., the first completion stage), the number j of generated (first category) data sets is first determined, and then clustering is performed iteratively according to the similarity between case samples (i.e., the first eigenvectors) until the number of iterations reaches a preset number (e.g., 10 times), and the multiple (first category data sets) obtained by the last iterative clustering are output, and the initial demand information is mapped to a vector with the same dimension as the first eigenvector (e.g., both are m-dimensional), and the first category data set (i.e., cluster) where the vector representing the initial demand information is located is determined based on the distance, and the initial demand information is supplemented using the average value of all case samples (i.e., the first eigenvector) of this cluster to obtain the (first) supplementary result (i.e., 1 completion sample) in this way. Still as Figure 4 As shown, in another supplementary processing scheme, the initial demand information is supplemented based on the probability classification method; first, the case sample (i.e., the first eigenvector) is randomly classified to generate multiple (second-class) data sets. In this embodiment, the number of the second-class data sets is the same as the number of the first-class data sets, both of which are j; next, the second-class data sets are iteratively updated according to the posterior probability of each first eigenvector in the case sample set until the change in the log-likelihood function value is less than the preset change (i.e., the predetermined threshold), and the second-class data set after the last iterative update is output for supplementary processing of the initial demand information; specifically, the posterior probability of the vector representing the initial demand information (i.e., the second eigenvector) under each second-class data set output after the last iterative update is determined, and the initial demand information is supplemented by the average value of all the first eigenvectors (i.e., other samples) contained in the second-class data set corresponding to the largest posterior probability value, and the (second) supplementary result (i.e., 2 supplementary samples) obtained based on the probability classification method is obtained. The final supplementary result (i.e., the target demand information) is the average value of multiple supplementary results (i.e., the first supplementary result and the second supplementary result) obtained by different supplementary processing methods.

[0072] Step S306, determining a cloud product combination corresponding to the initial demand information from a plurality of public cloud products provided by the public cloud according to the target demand information, and using the cloud product combination as a target cloud product solution, wherein the cloud product combination includes at least one public cloud product.

[0073] In step S306, the supplemented initial demand information (i.e., the target demand information) is parsed to obtain the precise cloud product resource configuration and business type contained therein; a single cloud product that meets the cloud product resource configuration and business type recorded in the target demand information is screened out from the cloud products provided by the public cloud. At this time, the above-mentioned single cloud product can be output as a cloud product solution; or, the cloud products provided by the public cloud are combined to obtain a cloud product combination that meets the cloud product resource configuration and business type recorded in the target demand information. This number of times, the cloud product combination will be output as a cloud product solution.

[0074] Figure 5 It is a structural diagram of an apparatus for determining a cloud product solution provided in accordance with an embodiment of the present application, comprising: an acquisition module 50, for acquiring initial demand information, wherein the initial demand information comprises: the business of applying the cloud product and the resource configuration of the cloud product; a supplementation module 52, for acquiring reference information upon receiving request information, and supplementing the initial demand information with the reference information to obtain target demand information, wherein the reference information comprises: a plurality of historical schemes that have been put into use, the historical business corresponding to each historical scheme, each historical scheme comprising the resource configuration of a plurality of target cloud products, and the request information is used to request the output of a cloud product solution that meets the initial demand information; a determination module 54, for determining a cloud product combination that meets the initial demand information from a plurality of public cloud products provided by a public cloud according to the target demand information, and outputting the cloud product combination as a cloud product solution, wherein the cloud product combination comprises at least one public cloud product.

[0075] Figure 6 It is a workflow diagram of the device that determines the cloud product solution, such as Figure 6As shown, the initial demand information is collected by the acquisition module 50, wherein the initial demand information is the user's demand for cloud products, including the business of the cloud product and the resource configuration of the cloud product; the business of the cloud product includes information such as business type (such as business field: finance, medical care, education); the resource configuration of the cloud product includes information such as demand for computing resources, demand for storage resources, demand for network resources, etc. If the information recorded in the acquired initial demand information covers accurate business information, such as the model of the cloud host, object storage requirements, elastic IP, shared bandwidth, and database, it is determined that there is no need to supplement the initial demand information, and the corresponding product solution is directly generated according to the initial demand information. If the initial demand information only records the business type, the nature and scale of the object (such as an enterprise or company) that applies the cloud product, and the numerical range of computing resources and the numerical range of network resources, it is determined that the initial demand information needs to be supplemented. In this embodiment, the initial demand information is supplemented by the supplement module 42, and the method provided in the embodiment of the present application only supplements the missing features in the initial demand information, and at the same time, the original information recorded in the initial demand information is retained. After the supplementation module 42 completes the supplementation process of the initial demand information, the determination module 54 determines a cloud product solution that meets the initial demand information according to the public cloud cloud product. In addition, the device for determining the cloud product solution can also determine and output a quotation (i.e., a second quotation) for the cloud product solution according to the unit price (i.e., the first quotation) and quantity of the cloud product (i.e., the target cloud product) included in the cloud product solution when receiving a request to output the cloud product solution.

[0076] It should be noted that Figure 5 The preferred implementation of the illustrated embodiment can be found in Figure 2 The relevant description of the illustrated embodiment will not be repeated here.

[0077] An embodiment of the present application also provides a non-volatile storage medium, in which a computer program is stored. The above method for determining a cloud product solution is executed by running the computer program on a device where the non-volatile storage medium is located.

[0078] The above-mentioned non-volatile storage medium is used to store programs that perform the following functions: obtain initial demand information, wherein the initial demand information includes: the business of applying cloud products and the resource configuration of cloud products; upon receiving request information, obtain reference information, and use the reference information to supplement the initial demand information to obtain target demand information, wherein the reference information includes: multiple historical plans that have been put into use, the historical business corresponding to each historical plan, and each historical plan contains the resource configuration of multiple target cloud products, and the request information is used to request the output of a cloud product solution that meets the initial demand information; determine a cloud product combination that meets the initial demand information among multiple public cloud products provided by the public cloud according to the target demand information, and output the cloud product combination as a cloud product solution, wherein the cloud product combination contains at least one public cloud product.

[0079] An embodiment of the present application also provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the above method for determining a cloud product solution through the computer program.

[0080] The processor in the above-mentioned electronic device is used to run a program that performs the following functions: obtaining initial demand information, wherein the initial demand information includes: the business of applying cloud products and the resource configuration of cloud products; upon receiving request information, obtaining reference information, and using the reference information to supplement the initial demand information to obtain target demand information, wherein the reference information includes: multiple historical plans that have been put into use, historical businesses corresponding to each historical plan, and each historical plan contains resource configurations of multiple target cloud products, and the request information is used to request the output of a cloud product solution that meets the initial demand information; according to the target demand information, determine a cloud product combination that meets the initial demand information among multiple public cloud products provided by the public cloud, and output the cloud product combination as a cloud product solution, wherein the cloud product combination contains at least one public cloud product.

[0081] An embodiment of the present application also provides a computer program product, including computer instructions, which implement the steps of the above method for determining a cloud product solution when executed by a processor.

[0082] It should be noted that the various modules in the device of the above-mentioned cloud product solution can be program modules (for example, a set of program instructions that implement a certain specific function) or hardware modules. For the latter, it can be expressed in the following forms, but is not limited to this: the expression form of each of the above-mentioned modules is a processor, or the functions of each of the above-mentioned modules are implemented by a processor.

[0083] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0084] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0085] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0086] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0087] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0088] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the relevant technology or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc. Various media that can store program codes.

[0089] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for determining a cloud product solution, characterized in that: include: Acquire initial demand information, wherein the initial demand information includes: the business of applying the cloud product and the resource configuration of the cloud product; Upon receiving the request information, obtain reference information, and use the reference information to supplement the initial demand information to obtain target demand information, wherein the reference information includes: multiple historical solutions that have been put into use, historical services corresponding to each of the historical solutions, and each of the historical solutions contains resource configurations of multiple target cloud products, and the request information is used to request the output of a cloud product solution that meets the initial demand information; A cloud product combination that meets the initial demand information is determined from a plurality of public cloud products provided by the public cloud according to the target demand information, and the cloud product combination is output as the cloud product solution, wherein the cloud product combination includes at least one of the public cloud products.

2. The method according to claim 1, characterized in that The initial demand information is supplemented by using the reference information to obtain target demand information, including: Determine the similarity between every two of the historical schemes, and generate a similarity matrix according to a plurality of the similarities; Generate multiple first feature vectors according to the similarity matrix, and perform cluster analysis on the multiple first feature vectors according to the distances between different first feature vectors to obtain multiple first-category data sets, wherein each first feature vector represents one of the historical plans, and each first-category data set contains multiple first feature vectors; supplement the initial demand information using the multiple first-category data sets to obtain a first supplement result; Classifying the first feature vector into a plurality of second-category data sets, wherein the number of the second-category data sets is the same as the number of the first-category data sets; and supplementing the initial demand information using the plurality of the second-category data sets to obtain a second supplementation result; An average value of the first supplement result and the second supplement result is determined as the target demand information.

3. The method according to claim 2, characterized in that Generating a plurality of first eigenvectors according to the similarity matrix includes: Determine the cumulative results of multiple similarities corresponding to each of the historical schemes, and generate a first matrix with the multiple cumulative results as elements, wherein the first matrix is ​​a diagonal matrix; Determining a difference between the first matrix and the similarity matrix as a second matrix, and performing normalization processing on the second matrix to obtain a normalized matrix, wherein a sum of multiple elements in each row of the normalized matrix is ​​a preset value; Performing eigenvalue decomposition processing on the normalized matrix to obtain a plurality of eigenvalues ​​and a plurality of first eigenvectors corresponding to the plurality of eigenvalues.

4. The method according to claim 2, characterized in that: Performing cluster analysis on the plurality of first feature vectors includes: Determine a plurality of target cluster centers among the plurality of the first feature vectors, wherein each of the target cluster centers is any one of the first feature vectors, the number of the target cluster centers is less than the number of the first feature vectors, and the number of the target cluster centers is determined by iteratively clustering the plurality of the first feature vectors; Determine the distance between each first remaining feature vector and each of the target cluster centers, and obtain a plurality of distances corresponding to each of the first remaining feature vectors, wherein the first remaining feature vectors are the feature vectors other than the target cluster center; For each of the first residual feature vectors, a first target distance with the smallest value is determined among the multiple distances corresponding to the first residual feature vector, and the first residual feature vectors and the target cluster centers corresponding to the first target distances are classified into one of the data sets, wherein each of the data sets contains one target cluster center and multiple first residual feature vectors.

5. The method according to claim 4, characterized in that The number of target cluster centers is determined by the following method: Iterative clustering is performed on the plurality of first feature vectors, wherein the following loop process is performed each time the iterative clustering is performed: Obtain a random number, and determine a target feature vector as an initial cluster center from a plurality of the first feature vectors according to the random number, wherein the target feature vector is any one of the first feature vectors, and the random number is used to indicate the number of the initial cluster centers; for each second remaining feature vector, determine a first initial cluster center having the smallest distance to the second remaining feature vector, and a second initial cluster center having a distance to the second remaining feature vector that is only greater than a distance between the second remaining feature vector and the first initial cluster center, wherein the second remaining feature vector is another first feature vector other than the initial cluster center; determine a first distance between the second remaining feature vector and each of the first feature vectors in the first data set to which the first initial cluster center belongs, and determine a first average value of a plurality of the first distances, and determine a second distance between the second remaining feature vector and each of the first feature vectors in the second data set to which the second initial cluster center belongs, and determine a second average value of a plurality of the second distances; determine a first silhouette coefficient for each of the second remaining feature vectors according to the first average value and the second average value, and determine the average value of the plurality of the first silhouette coefficients as the second silhouette coefficient corresponding to the random number; When the number of iterative clustering reaches a preset number, a target silhouette coefficient with the largest value is determined among a plurality of second silhouette coefficients corresponding to a plurality of the random numbers, and the target silhouette coefficient is determined as the number of the target cluster centers.

6. The method according to claim 4, characterized in that The initial demand information is supplemented by using the plurality of data sets to obtain a first supplement result, including: Mapping the initial requirement information into a second feature vector having the same dimension as the first feature vector, wherein the second feature vector includes a first target element whose mapping value is empty; Determine the data set to which the target cluster center closest to the second feature vector belongs among the plurality of target cluster centers as the target data set; Determine a target position of the first target element in the second feature vector, determine a second target element indicated by each of the first feature vectors in the target data set at the target position, and determine an average value of a plurality of the second target elements as a first supplementary element; The first supplementary element is filled into the target position of the second feature vector to obtain the first supplementary result.

7. The method according to claim 6, characterized in that The initial demand information is supplemented by using a plurality of the second-category data sets to obtain a second supplement result, including: Determine the posterior probability of the second eigenvector in each of the second-category data sets, and determine the second-category data set corresponding to the posterior probability with the largest value as the third-category data set; Determine an average value of all the first feature vectors in the third type of data set to obtain an average feature vector; The element indicated by the target position in the average feature vector is determined as a second supplementary element, and the second supplementary element is supplemented to the target position of the second feature vector to obtain the second supplementary result.

8. The method according to claim 2, characterized in that: Classifying the first feature vector into a plurality of second-category data sets, including: Randomly classify the first feature vectors into a plurality of initial data sets, wherein the number of the initial data sets is the same as the number of the first type of data sets; Iterative classification is performed on a plurality of the initial data sets, wherein each time the iterative classification is performed, the following loop process is performed: for each of the initial data sets, an initial mean vector, an initial covariance matrix and an initial weight coefficient of the initial data set are determined, wherein the initial mean vector is an average value of a plurality of the first eigenvectors contained in the initial data set, the elements in the initial covariance matrix include: a variance determined according to the initial data set, a covariance determined according to the initial data set, and the initial weight coefficient is used to indicate the number of the first eigenvectors contained in the initial data set; Determine the posterior probability of each first eigenvector in each of the initial data sets according to the initial mean vector, the initial covariance matrix and the initial weight coefficient of each of the initial data sets, wherein the posterior probability is used to indicate the probability that the first eigenvector is classified into the initial data set; Determine a target posterior probability with the largest value among the multiple posterior probabilities corresponding to the first feature vector, and reclassify the first feature vector into the initial data set corresponding to the target posterior probability to obtain an update result of the initial data set; When the difference between the first function value and the second function value is less than a preset value, the iterative classification is determined to be completed, and the multiple updated initial data sets output by the last iterative classification are determined as the second type of data sets, wherein the first function value is a function value determined based on the initial mean vector, initial covariance matrix and initial weight coefficient determined in the last iterative classification, and the second function value is a function value determined based on the initial mean vector, initial covariance matrix and initial weight coefficient generated in the next iterative classification.

9. The method according to claim 1, characterized in that: The method further comprises: Upon receiving the quotation request, determining each target cloud product included in the cloud product solution, and determining a first quotation for each target cloud product; A second quotation for the cloud product solution is determined according to the quotation for the target cloud product, and the second quotation is output.

10. A method for determining a cloud product solution, characterized in that: include: Acquire initial demand information, wherein the initial demand information includes: the business of applying the cloud product and the resource configuration of the cloud product; Acquire reference information, and use the reference information to supplement the initial demand information to obtain target demand information, wherein the reference information includes: multiple historical solutions that have been put into use, historical businesses corresponding to each of the historical solutions, and resource configurations of multiple target cloud products included in each of the historical solutions; According to the target demand information, a cloud product combination corresponding to the initial demand information is determined from a plurality of public cloud cloud products provided by the public cloud, and the cloud product combination is used as a target cloud product solution, wherein the cloud product combination includes at least one of the public cloud cloud products.

11. A device for determining a cloud product solution, characterized in that: include: An acquisition module is used to acquire initial demand information, wherein the initial demand information includes: the business of the application cloud product and the resource configuration of the cloud product; A supplementation module is used to obtain reference information when receiving the request information, and use the reference information to supplement the initial demand information to obtain target demand information, wherein the reference information includes: multiple historical solutions that have been put into use, historical businesses corresponding to each of the historical solutions, and each of the historical solutions contains resource configurations of multiple target cloud products, and the request information is used to request the output of a cloud product solution that meets the initial demand information; A determination module is used to determine a cloud product combination that meets the initial demand information from a plurality of public cloud products provided by the public cloud according to the target demand information, and output the cloud product combination as the cloud product solution, wherein the cloud product combination includes at least one of the public cloud products.

12. A non-volatile storage medium, characterized in that: A computer program is stored in the non-volatile storage medium, wherein the method for determining a cloud product solution as described in any one of claims 1 to 9 and claim 10 is executed by running the computer program on the device where the non-volatile storage medium is located.

13. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method for determining a cloud product solution as described in any one of claims 1 to 9 and claim 10 through the computer program.

14. A computer program product comprising computer instructions, characterized in that When the computer instructions are executed by a processor, the steps of the method for determining a cloud product solution described in any one of claims 1 to 9 and claim 10 are implemented.

Images