Customer satisfaction prediction method, device, equipment and medium

Abstract

The present disclosure provides a method for predicting customer satisfaction, which can be used in the field of big data technology. The method includes: obtaining personal information of n customers, wherein the personal information of each customer includes N information items; quantifying the personal information of n customers to obtain a personal information matrix; for the personal information matrix, using the spectral clustering method to calculate the graph Laplacian matrix corresponding to the personal information matrix; calculating the spectral density of the graph Laplacian matrix; constructing a graph filter according to the graph Laplacian matrix, wherein in the process of constructing the graph filter, the passband interval of the graph filter is determined according to the spectral density; generating a feature matrix according to the graph filter; based on the feature matrix, using the clustering method to classify the n customers; and predicting the satisfaction of at least one customer according to the classified n customers. According to the prediction method in the embodiment of the present disclosure, the amount of calculation, the complexity of calculation and the memory usage are greatly reduced.

Description

Technical Field

[0001] The present disclosure relates to the field of big data technology, and in particular to a method, device, equipment and medium for predicting customer satisfaction. Background Art

[0002] Customer satisfaction is one of the important indicators for evaluating the quality of service in the service industry. At present, the evaluation and prediction of customer satisfaction mainly adopts statistical methods or clustering methods. In the statistical method, by collecting user information and satisfaction, the distribution of satisfaction of users with certain common characteristics is statistically calculated. The use of statistical methods to predict satisfaction has the problem of low efficiency and poor prediction accuracy. In the clustering method, the big data analysis of customer satisfaction usually takes the information of each customer as a sample, classifies these customers using machine learning methods, obtains the relationship between customer information and satisfaction information, and then predicts customer satisfaction. However, for financial companies with more than 10 million customer data, every time a new customer or a new information is added, the machine needs to recalculate and classify all the data, which will generate a huge amount of calculation, time-consuming and resource-consuming. Summary of the invention

[0003] In view of the above problems, the present disclosure provides a method, apparatus, device, medium and program product for predicting customer satisfaction.

[0004] According to a first aspect of the present disclosure, a method for predicting customer satisfaction is provided, comprising the following steps:

[0005] Obtain the customer's authorization to obtain personal information;

[0006] Obtaining personal information of n customers with the authorization of the customer, wherein the personal information of each customer includes N information items, and the N information items are all related to customer satisfaction, where n is an integer greater than or equal to 1, and N is an integer greater than or equal to 2;

[0007] Quantifying the personal information of the n customers to obtain a personal information matrix, wherein the personal information matrix is ​​a matrix with n rows and N columns, and each row of the personal information matrix represents the quantified personal information of a customer;

[0008] For the personal information matrix, a graph Laplacian matrix corresponding to the personal information matrix is ​​calculated using a spectral clustering method, wherein the graph Laplacian matrix is ​​a matrix with n rows and n columns;

[0009] Calculating the spectral density of the graph Laplacian matrix;

[0010] constructing a graph filter according to the graph Laplacian matrix, wherein, in the process of constructing the graph filter, a passband interval of the graph filter is determined according to the spectral density;

[0011] Generate a feature matrix according to the graph filter, wherein the feature matrix is ​​a matrix with n rows and b columns, b is a positive integer and 1≤b<n;

[0012] Based on the feature matrix, classifying the n customers using a clustering method; and

[0013] Based on the classified n customers, the satisfaction level of at least one of the n customers is predicted.

[0014] According to some exemplary embodiments, determining the passband interval of the graph filter according to the spectral density specifically includes: using an iterative method to calculate the eigenvalues ​​of the graph Laplacian matrix according to the spectral density; and determining the passband interval of the graph filter according to the eigenvalues.

[0015] According to some exemplary embodiments, the adopting an iterative method to calculate the eigenvalues ​​of the graph Laplacian matrix according to the spectral density specifically includes: iteratively calculating the upper limit of the interval of the numerical integral of the spectral density according to a specified integration step; and when the numerical integral of the spectral density is greater than or equal to b, determining the upper limit of the interval of the numerical integral of the spectral density as the bth eigenvalue of the graph Laplacian matrix.

[0016] According to some exemplary embodiments, the using an iterative method to calculate the eigenvalues ​​of the graph Laplacian matrix according to the spectral density further specifically includes: when the iterative calculation is completed, determining the upper limit of the interval of the numerical integration of the spectral density as the b+1th eigenvalue of the graph Laplacian matrix.

[0017] According to some exemplary embodiments, determining the passband range of the graph filter based on the eigenvalues ​​specifically includes: determining an average value of the bth eigenvalue and the b+1th eigenvalue as an upper limit of the passband range of the graph filter.

[0018] According to some exemplary embodiments, calculating the spectral density of the graph Laplacian matrix specifically includes: calculating the graph Laplacian matrix polynomial according to the graph Laplacian matrix; calculating the quadratic form of the graph Laplacian matrix polynomial; and calculating the spectral density of the graph Laplacian matrix according to the quadratic form of the graph Laplacian matrix polynomial.

[0019] According to some exemplary embodiments, constructing the graph filter according to the graph Laplacian matrix specifically includes: constructing the graph filter according to the graph Laplacian matrix polynomial and the passband interval of the graph filter.

[0020] According to a second aspect of the present disclosure, a device for predicting customer satisfaction is provided, comprising: a customer authorization acquisition module, used to obtain the customer's authorization to obtain personal information; a personal information acquisition module, used to: obtain the personal information of n customers when the customer's authorization to obtain personal information is obtained, wherein the personal information of each customer includes N information items, and the N information items are all related to customer satisfaction, n is an integer greater than or equal to 1, and N is an integer greater than or equal to 2; a personal information matrix acquisition module, used to: quantify the personal information of the n customers to obtain a personal information matrix, wherein the personal information matrix is ​​a matrix with n rows and N columns, and each row of the personal information matrix represents the quantified personal information of a customer; a graph Laplacian matrix calculation module, used to: calculate the personal information matrix using the spectrum A clustering method is used to calculate a graph Laplacian matrix corresponding to the personal information matrix, wherein the graph Laplacian matrix is ​​a matrix with n rows and n columns; a spectral density calculation module is used to calculate the spectral density of the graph Laplacian matrix; a graph filter construction module is used to construct a graph filter according to the graph Laplacian matrix, wherein in the process of constructing the graph filter, the passband interval of the graph filter is determined according to the spectral density; a feature matrix acquisition module is used to generate a feature matrix according to the graph filter, wherein the feature matrix is ​​a matrix with n rows and b columns, b is a positive integer and 1≤b<n; a classification module is used to classify the n customers based on the feature matrix using a clustering method; and a customer satisfaction prediction module is used to predict the satisfaction of at least one of the n customers according to the classified n customers.

[0021] A third aspect of the present disclosure provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the above method.

[0022] The fourth aspect of the present disclosure also provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the above method.

[0023] The fifth aspect of the present disclosure also provides a computer program product, including a computer program, which implements the above method when executed by a processor. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0025] Figure 1A diagram schematically illustrates an application scenario of a method for predicting customer satisfaction according to an embodiment of the present disclosure;

[0026] Figure 2 A flowchart schematically shows a method for predicting customer satisfaction according to an embodiment of the present disclosure;

[0027] Figure 3 is a schematic flow chart of calculating the spectral density of a graph Laplacian matrix according to an embodiment of the present disclosure;

[0028] Figure 4 is a schematic flow chart of constructing a graph filter according to an embodiment of the present disclosure;

[0029] Figure 5 A structural block diagram schematically shows a device for predicting customer satisfaction according to an embodiment of the present disclosure; and

[0030] Figure 6 A block diagram of an electronic device suitable for implementing a method for predicting customer satisfaction according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0031] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0032] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0033] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0034] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0035] Customer satisfaction is one of the important indicators for evaluating business quality in the service industry. Currently, statistical methods or clustering methods are mainly used to evaluate and predict customer satisfaction.

[0036] In the statistical method, by collecting user information and satisfaction, the distribution of satisfaction of users with certain common characteristics is statistically analyzed. For example, by statistically analyzing the satisfaction of female users, 11.318% are very dissatisfied, 20.888% are generally satisfied, 21.666% are very satisfied...; by statistically analyzing the satisfaction of users aged 20-25, 2.318% are very dissatisfied, 40.888% are generally satisfied, and 31.666% are very satisfied... Using statistical methods to predict satisfaction has the problem of low efficiency and poor prediction accuracy. For example, it is entirely doubtful that there are great differences within the group of female users, and their satisfaction does not have commonality. Therefore, statistically analyzing the satisfaction of female users does not have a particularly strong guiding significance for decision-making.

[0037] In the clustering method, by collecting customer information and satisfaction information, each information is used as a sample, and the sample is processed by a machine learning method (such as SVM, neural network, decision tree, etc.), the relationship between customer information and satisfaction information is obtained, and the customer satisfaction is subsequently predicted. The inventor found that when the sample is processed by the machine learning method, the information is retained intact, and the amount of calculation is too large. For example, for the data of 100,000 users, a dense matrix of 100,000 orders will be formed. This matrix size is a heavy burden for the vast majority of current supercomputer systems: it usually requires dozens to thousands of computing nodes and a long time of calculation. Moreover, for financial companies with more than 10 million customer data, every time a new customer is added or a new information is added, the machine needs to recalculate and classify all the data, which will generate a huge amount of calculation, time-consuming and resource-consuming.

[0038] In order to facilitate understanding of the technical solution of the present application, the technical terms involved in the present application will be introduced below.

[0039] Spectral clustering algorithm: an unsupervised machine learning method. The spectral clustering algorithm is based on the spectral graph theory in graph theory. Its essence is to transform the clustering problem into the optimal partitioning problem of the graph. Compared with traditional clustering algorithms, it has the advantages of being able to cluster on sample spaces of arbitrary shapes and converge to the global optimal solution.

[0040] Dimensionality reduction: It is a key step in the spectral clustering algorithm. In this application, the input data can be reduced to reduce the amount of calculation. For example, the amount of data is reduced from n×n to n×b, where b must be less than n.

[0041] Eigenvector and eigenvalue: The eigenvector of a matrix is ​​one of the important concepts in matrix theory. The eigenvector of a linear transformation is a non-degenerate vector whose direction remains unchanged under the transformation. The scale of the vector under this transformation becomes its eigenvalue. Mathematically, if the vector v and the transformation A satisfy Av=λv, then the vector v is called an eigenvector of the transformation A, and λ is the corresponding eigenvalue.

[0042] Feature space: The space where the feature vectors are located. Each feature corresponds to a unique coordinate in the feature space.

[0043] Graph: A graph is a data structure consisting of points and lines. A graph Laplacian matrix is ​​a square matrix used to describe a graph. A graph signal is a vector that reflects the effect of the graph Laplacian matrix on the vector.

[0044] Graph signal: Graph signal is a data representation form that describes unstructured data. The processing object of classical digital signal processing is time signal, which is composed of a series of sampling points at different times. This implies a very regular data structure, that is, each sampling point is arranged in a regular time sequence. However, in practice, we often need to process irregularly structured data, which includes not only the numerical information itself, but also the structural association information between the data. For example, traffic networks, wireless sensor networks, social networks, etc. Graph signals break the regular structure limitations of classical digital signal processing and are dedicated to processing signals on graph structures.

[0045] The following will describe embodiments of the present application with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the disclosure scope of the present application. In the following detailed description, for ease of explanation, many specific details are set forth and a comprehensive explanation of the embodiments of the present application is provided. However, one or more embodiments may also be implemented without these specific details. In addition, in the following description, the description of known structures and technologies is omitted to avoid unnecessary confusion.

[0046] It should be noted that in the technical solution of this application, the acquisition, storage and application of customer personal information involved are in compliance with the provisions of relevant laws and regulations, necessary confidentiality measures have been taken, and do not violate public order and good morals.

[0047] Figure 1 The application scenario diagram of the customer satisfaction prediction method according to the embodiment of the present disclosure is schematically shown.

[0048] like Figure 1 As shown, the application scenario 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0049] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only examples).

[0050] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0051] The server 105 may be a server that provides various services, such as a background management server (only an example) that provides support for websites browsed by users using the terminal devices 101, 102, and 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0052] It should be noted that the customer satisfaction prediction method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the customer satisfaction prediction device provided in the embodiment of the present disclosure can generally be set in the server 105. The customer satisfaction prediction method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the customer satisfaction prediction device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105.

[0053] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0054] Figure 2 The flowchart of the method for predicting customer satisfaction according to the embodiment of the present disclosure is schematically shown.

[0055] like Figure 2 As shown, the prediction method of this embodiment includes operations S110 to S190.

[0056] In operation S110, the customer's authorization to obtain personal information is obtained.

[0057] In the embodiments of the present disclosure, before obtaining the information of the customer, the customer's consent or authorization is required. For example, before operation S120, the customer may be notified and a request may be issued to the customer to obtain the personal information and other associated information. When the customer agrees or authorizes the acquisition of the personal information, operation S120 is performed.

[0058] In operation S120, with the customer's authorization to obtain personal information, personal information of n customers is obtained, where the personal information of each customer includes N information items, and the N information items are all related to customer satisfaction, n is an integer greater than or equal to 1, and N is an integer greater than or equal to 2.

[0059] Before predicting customer satisfaction, it is necessary to obtain the customer's personal information, commonly known as customer information. From the customer information, we can usually learn about the customer's customer management ability, profitability, debt repayment ability, development ability, customer quality and credit status, and judge the customer's credit based on this relevant information.

[0060] The channels for obtaining a customer's personal information may be diverse. For example, in a specific example, the following text segments may be included: "height, weight, age, income, hobbies, savings, education level, occupation, ...".

[0061] After obtaining the customer's personal information, organize the personal information into N information items, that is, each customer corresponds to N information items. However, since some of the N information items may be related, for example, if the company or individual has high income, the tax will also increase accordingly, and the two information items of income information and tax amount must exist in direct proportion, the related information items can be merged to achieve the purpose of dimensionality reduction. This application involves dimensionality reduction, so in this operation, the information item N obtained is an integer greater than or equal to 2.

[0062] It is understandable that when there is a negative ratio between multiple information items, they can also be merged and reduced in dimension.

[0063] In operation S130, the personal information of n customers is quantified to obtain a personal information matrix, wherein the personal information matrix is ​​a matrix of n rows and N columns, and each row of the personal information matrix represents the quantified personal information of one customer.

[0064] Before dimensionality reduction, the personal information of n customers is mathematically processed.

[0065] The specific processing method can be adjusted according to the specific content of the information item. For example, profitability can be expressed in numbers, including but not limited to fixed salary, profit, gross profit margin, etc.

[0066] Using the quantified customer data, we can use element X ij In other words, i represents the i-th customer and j represents the j-th information item of the i-th customer. Therefore, both i and j can be represented as nodes in an undirected graph.

[0067] In operation S140, a graph Laplacian matrix corresponding to the personal information matrix is ​​calculated using a spectral clustering method, wherein the graph Laplacian matrix is ​​a matrix with n rows and n columns.

[0068] When calculating the graph Laplace, it is necessary to first use the quantized personal information to generate an adjacency matrix W. The adjacency matrix is ​​a matrix representation of a graph. With it, the structure of the graph can be easily stored and the problem of the graph can be studied using linear algebra. In this application, the solution is for n customers, so the adjacency matrix W is an n*n matrix, that is, the adjacency matrix W is an n*n dimensional matrix, where the matrix elements are the X obtained in operation S130. ij The element is expressed as the weight of the edge (i, j). If there is no edge connection between two nodes, the corresponding element in the adjacency matrix is ​​0. Specifically, the KNN algorithm is used to traverse all sample points and only the nearest k points of each sample are retained as neighbors, that is, only the W between the k points closest to the sample is ij >0, and the rest are set to 0.

[0069] The calculation formula is as follows:

[0070]

[0071]

[0072] In the above formula, X i is the data of the ith row, i.e., the personal information of the ith customer, X j The data in the j-th row is the personal information of the j-th customer.

[0073] It can be understood that formula (1) represents that as long as X i In X j In the K neighboring set of ij ; (2) Formula represents X i and X j The two points must be in each other's k-neighboring set to retain W ij .

[0074] For W ij The calculation of the Euclidean distance between any two points (X i and X j ) is a specific embodiment of the distance between. In the process of presetting the function, polynomial kernel function, Gaussian kernel function and Sigmoid kernel function are commonly used.

[0075] As for this application, the most commonly used Gaussian kernel function is used for calculation.

[0076]

[0077] After obtaining the adjacency matrix W through the above method, the diagonal matrix D, that is, the degree matrix, is further calculated.

[0078] Since the present application belongs to an undirected graph, for an undirected graph, the weighted degree of a node is the sum of the weight values ​​of all edges associated with the node. The weighted degree of node i in the adjacency matrix W of the undirected graph is the sum of the elements in the i-th row of the adjacency matrix.

[0079] According to the adjacent matrix W, the corresponding diagonal matrix D is calculated as follows:

[0080]

[0081] Finally, the diagonal matrix D is:

[0082]

[0083] According to the n nodes (n customers) in the undirected graph, the adjacency matrix is ​​W and the diagonal matrix is ​​D. The graph Laplacian matrix can be defined based on the adjacency matrix W and the diagonal matrix D. That is, the graph Laplacian matrix A is defined as the difference between the diagonal matrix D and the adjacency matrix W:

[0084] A=DW

[0085] In operation S150, the spectral density of the graph Laplacian matrix A is calculated.

[0086] In conventional signal processing, spectral density refers to the root mean square value per unit bandwidth when the bandwidth of the signal approaches zero. In graph signal processing, the spectral density of a graph signal can be understood as the distribution characteristics of the graph signal at each frequency component. For example, the spectral density here can be understood as the distribution characteristics of the eigenvalues ​​of the graph Laplacian matrix.

[0087] Correspondingly, in this article, the expression “image filtering” can be understood as an operation of enhancing or attenuating the intensity of each frequency component in the frequency spectrum of a given image signal.

[0088] Figure 3 is a schematic flow chart of calculating the spectral density of the graph Laplacian matrix according to an embodiment of the present disclosure. Figure 3 In an embodiment of the present disclosure, the operation S150 may include sub-operations S1510 to S1530.

[0089] In sub-operation S1510, the graph Laplacian matrix polynomial is calculated according to the graph Laplacian matrix.

[0090] In sub-operation S1520, a quadratic form of the graph Laplacian matrix polynomial is calculated.

[0091] In sub-operation S1530, a spectral density of the graph Laplacian matrix is ​​calculated according to the quadratic form of the graph Laplacian matrix polynomial.

[0092] For example, in some exemplary embodiments of the present disclosure, the matrix polynomial of the graph Laplacian matrix A may be calculated using Chebyshev polynomials, and specifically, reference may be made to the following formula (3).

[0093] C i (A) = cos(i*arccos(A)) (3)

[0094] Wherein i=1, 2, ..., m, where m is the highest degree of the Chebyshev polynomial.

[0095] Then, the quadratic form of the graph Laplacian matrix polynomial is calculated using formula (4).

[0096]

[0097] where v1, v2, ..., v b are b n-dimensional random vectors, for example, the elements of each vector satisfy the Gaussian distribution.

[0098] Then, according to the quadratic form of the graph Laplacian matrix polynomial, the spectral density φ of the graph Laplacian matrix is ​​calculated using formula (5): m (t).

[0099]

[0100] Among them, the coefficient g i It can be determined by formula (6):

[0101]

[0102] Among them, i=0, 1, 2, ..., m.

[0103] In operation S160, a graph filter is constructed according to the graph Laplacian matrix, wherein in the process of constructing the graph filter, a passband interval of the graph filter is determined according to the spectral density.

[0104] Figure 4 is a schematic flow chart of constructing a graph filter according to an embodiment of the present disclosure. Figure 4 In an embodiment of the present disclosure, the operation S160 may include sub-operations S1610 to S1620.

[0105] In sub-operation S1610, an iterative method is used to calculate the eigenvalues ​​of the graph Laplacian matrix according to the spectral density.

[0106] For example, the adopting an iterative method to calculate the eigenvalues ​​of the graph Laplacian matrix according to the spectral density specifically includes: iteratively calculating the upper limit of the interval of the numerical integral of the spectral density according to a specified integration step; and when the numerical integral of the spectral density is greater than or equal to b, determining the upper limit of the interval of the numerical integral of the spectral density as the bth eigenvalue of the graph Laplacian matrix.

[0107] For example, the iterative method is used to calculate the eigenvalues ​​of the graph Laplacian matrix according to the spectral density, and specifically includes: when the iterative calculation is completed, determining the upper limit of the interval of the numerical integration of the spectral density as the b+1th eigenvalue of the graph Laplacian matrix.

[0108] In sub-operation S1620, a passband interval of the graph filter is determined according to the eigenvalue.

[0109] For example, determining the passband interval of the graph filter according to the eigenvalue specifically includes: determining an average value of the bth eigenvalue and the b+1th eigenvalue as an upper limit of the passband interval of the graph filter.

[0110] Specifically, in the embodiments of the present disclosure, a graph filter may be constructed according to the following operations.

[0111] First, the spectral density φ of the graph Laplacian matrix is ​​calculated using the above method. m (t).

[0112] Then, the initial values ​​of the numerical integral N of the spectral density and the upper limit λ of the interval of the numerical integral of the spectral density are given, for example, N=0, λ=-1.

[0113] Then, the following iterative calculation process is performed until the b+1th eigenvalue is determined:

[0114] λ=λ+h

[0115] N=N+n·h·φ m (λ)

[0116] When N ≥ b + 1,

[0117] When N ≥ b, λ b =λ,

[0118] At the end of the iteration process,

[0119] In the above iterative calculation process, λ is a continuously growing variable, which represents the upper limit of the interval of the current spectral density numerical integration. When the numerical integration N is b, λ is the upper limit of the interval where the bth eigenvalue is located, and the λ at this time is used to approximately replace the bth eigenvalue; when the algorithm ends, λ is the upper limit of the interval where the b+1th eigenvalue is located, and the λ at this time is used to replace the b+1th eigenvalue. The middle position between the bth and b+1th eigenvalues ​​is the passband boundary position to be set for the graph filter.

[0120] In an embodiment of the present disclosure, the graph filter is constructed according to the graph Laplacian matrix polynomial and the passband interval of the graph filter.

[0121] For example, the graph filter can be constructed according to formula (7):

[0122]

[0123] in,

[0124] [a, b] is the passband interval of the image filter, for example, a=0,

[0125] In operation S170 , a feature matrix is ​​generated according to the graph filter, wherein the feature matrix is ​​a matrix with n rows and b columns, b is a positive integer and 1≤b<n.

[0126] For example, the feature matrix Y can be generated by the graph filter according to the following formula (8):

[0127] Y=P[v1,v2,...,v b ] (8)

[0128] In formula (8), P is an n-order square matrix, [v1, v2, ..., v b ] is an n-row and b-column matrix composed of random vectors, and Y is an n-row and b-column matrix.

[0129] In operation S180, the n customers are classified using a clustering method based on the feature matrix.

[0130] For the n*b-dimensional feature matrix P, use the data in each row as a sample and use K-means to classify customers. It should be clear that the feature matrix P is a matrix with n rows and b columns. This operation is to use the dimensionality-reduced personal information b corresponding to each customer in the customer n as a sample for classification.

[0131] In an embodiment of the present disclosure, after using the dimensionality reduction method as described above to obtain an n*b-dimensional feature matrix, the data in each row is used as a sample to classify the customers. The embodiment of the present disclosure is not limited to the above-mentioned K-means clustering method, and a trained neural network or decision tree (including the classic decision tree method and derivative methods such as random forest) can also be used to classify all customers.

[0132] It should be noted that in practical applications, experience shows that the final b column should be the same as the number of categories after the customers are classified, otherwise the error will become larger. That is, if the customers are to be classified into 7 categories, the feature matrix P should have n*7 data, that is, the 7 eigenvectors of the graph Laplace A are obtained, and the feature matrix P after dimensionality reduction is a matrix with n rows and 7 columns.

[0133] In operation S190, customer satisfaction of at least one of the n customers is predicted based on the classified n customers.

[0134] For example, the distribution of satisfaction of customers in the i-th category (i is a number between 1 and b) is the probability distribution of the satisfaction of a certain customer, Li Si, among n customers. For example, in the i-th category, 11.667% are very dissatisfied, 15.86% are dissatisfied, 23.11% are satisfied... When Li Si is assigned to the i-th category, the conclusion is: the probability of Li Si's satisfaction is 11.667% very dissatisfied, 15.86% dissatisfied, 23.11% satisfied...

[0135] According to the prediction method in the embodiment of the present disclosure, the characteristic of spectral clustering dimensionality reduction is used to reduce the dimension of the data (from n*n dimension to n*b). The dimensionality reduction process is different from other spectral clustering methods, and the spectral density is used to construct the graph filter. Other spectral clustering methods construct filters by solving eigenvalues, and the computational complexity of solving eigenvalues ​​is greater than that of the embodiment of the present disclosure, and the memory access complexity is also greater than that of the embodiment of the present disclosure. In other words, the method provided by the embodiment of the present disclosure has a fast calculation speed and occupies little memory, which greatly reduces the amount of calculation, computational complexity and memory usage.

[0136] Based on the above customer satisfaction prediction method, the present disclosure also provides a customer satisfaction prediction device. Figure 5 The device is described in detail. Figure 5 The structural block diagram of the prediction device according to an embodiment of the present disclosure is schematically shown.

[0137] like Figure 5 As shown, the prediction device 800 of this embodiment includes a customer authorization acquisition module 810, a personal information acquisition module 820, a personal information matrix acquisition module 830, a graph Laplacian matrix calculation module 840, a spectral density calculation module 850, a graph filter construction module 860, a feature matrix acquisition module 870, a classification module 880 and a customer satisfaction prediction module 890.

[0138] The client authorization acquisition module 810 is used to obtain the client's authorization to obtain personal information. In one embodiment, the client authorization acquisition module 810 can be used to perform the operation S110 described above, which will not be described in detail here.

[0139] The personal information acquisition module 820 is used to obtain the personal information of n customers, where the personal information of each customer includes N information items, and the N information items are all related to customer satisfaction, n is an integer greater than or equal to 1, and N is an integer greater than or equal to 2. In one embodiment, the personal information acquisition module 820 can be used to perform the operation S120 described above, which will not be repeated here.

[0140] The personal information matrix acquisition module 830 is used to quantify the personal information of n customers to obtain a personal information matrix, wherein the personal information matrix is ​​a matrix of n rows and N columns, and each row of the personal information matrix represents the quantified personal information of a customer. In one embodiment, the personal information matrix acquisition module can be used to perform the operation S130 described above, which will not be repeated here.

[0141] The graph Laplacian matrix calculation module 840 is used to calculate the graph Laplacian matrix corresponding to the personal information matrix using the spectral clustering method, wherein the graph Laplacian matrix is ​​a matrix with n rows and n columns. In one embodiment, the graph Laplacian matrix calculation module can be used to perform the operation S140 described above, which will not be repeated here.

[0142] The spectral density calculation module 850 is used to calculate the spectral density of the graph Laplacian matrix. In one embodiment, the spectral density calculation module can be used to perform the operation S150 described above, which will not be described in detail here.

[0143] The graph filter construction module 860 is used to: construct a graph filter according to the graph Laplacian matrix, wherein, in the process of constructing the graph filter, the passband interval of the graph filter is determined according to the spectral density. In one embodiment, the graph filter construction module can be used to perform the operation S160 described above, which will not be repeated here.

[0144] The feature matrix acquisition module 870 is used to generate a feature matrix according to the graph filter, wherein the feature matrix is ​​a matrix with n rows and b columns, b is a positive integer and 1≤b<n. In one embodiment, the feature matrix acquisition module can be used to perform the operation S170 described above, which will not be repeated here.

[0145] The classification module 880 is used to classify the n customers using a clustering method based on the feature matrix. In one embodiment, the classification module can be used to perform the operation S180 described above, which will not be described in detail here.

[0146] The customer satisfaction prediction module 890 is used to predict the customer satisfaction of at least one of the n customers according to the classified n customers. In one embodiment, the customer satisfaction prediction module 890 can be used to perform the operation S190 described above, which will not be described in detail here.

[0147] According to the prediction device in the embodiment of the present disclosure, the above prediction method can be executed, and the characteristics of spectral clustering dimensionality reduction are used to reduce the dimension of the data (from n*n dimensions to n*b). The dimensionality reduction process is different from other spectral clustering methods. The spectral density is used to construct the graph filter, which can greatly reduce the amount of calculation, computational complexity and memory usage.

[0148] According to an embodiment of the present disclosure, any multiple modules among the customer authorization acquisition module 810, the personal information acquisition module 820, the personal information matrix acquisition module 830, the graph Laplacian matrix calculation module 840, the spectral density calculation module 850, the graph filter construction module 860, the feature matrix acquisition module 870, the classification module 880 and the customer satisfaction prediction module 890 can be combined into one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the customer authorization acquisition module 810, the personal information acquisition module 820, the personal information matrix acquisition module 830, the graph Laplacian matrix calculation module 840, the spectral density calculation module 850, the graph filter construction module 860, the feature matrix acquisition module 870, the classification module 880 and the customer satisfaction prediction module 890 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or can be implemented in any one of the three implementation methods of software, hardware and firmware or in an appropriate combination of any of them. Alternatively, at least one of the customer authorization acquisition module 810, the personal information acquisition module 820, the personal information matrix acquisition module 830, the graph Laplacian matrix calculation module 840, the spectral density calculation module 850, the graph filter construction module 860, the feature matrix acquisition module 870, the classification module 880 and the customer satisfaction prediction module 890 can be at least partially implemented as a computer program module, which can perform the corresponding function when the computer program module is executed.

[0149] Figure 6 A block diagram of an electronic device suitable for implementing a method for predicting customer satisfaction according to an embodiment of the present disclosure is schematically shown.

[0150] like Figure 6As shown, the electronic device 900 according to an embodiment of the present disclosure includes a processor 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 902 or a program loaded from a storage part 908 into a random access memory (RAM) 903. The processor 901 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 901 may also include an onboard memory for caching purposes. The processor 901 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0151] In RAM 903, various programs and data required for the operation of electronic device 900 are stored. Processor 901, ROM 902 and RAM 903 are connected to each other via bus 904. Processor 901 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 902 and / or RAM 903. It should be noted that the program can also be stored in one or more memories other than ROM 902 and RAM 903. Processor 901 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in the one or more memories.

[0152] According to an embodiment of the present disclosure, the electronic device 900 may further include an input / output (I / O) interface 905, which is also connected to the bus 904. The electronic device 900 may further include one or more of the following components connected to the I / O interface 905: an input portion 906 including a keyboard, a mouse, etc.; an output portion 907 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 908 including a hard disk, etc.; and a communication portion 909 including a network interface card such as a LAN card, a modem, etc. The communication portion 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to the I / O interface 905 as needed. A removable medium 911, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 910 as needed, so that a computer program read therefrom is installed into the storage portion 908 as needed.

[0153] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0154] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present disclosure, a computer-readable storage medium may include the ROM 902 and / or RAM 903 described above and / or one or more memories other than ROM 902 and RAM 903.

[0155] The embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the customer satisfaction prediction method provided by the embodiment of the present disclosure.

[0156] The above functions defined in the system / device of the embodiment of the present disclosure are performed when the computer program is executed by the processor 901. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0157] In one embodiment, the computer program may be based on a tangible storage medium such as an optical storage device, a magnetic storage device, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of a signal on a network medium, and downloaded and installed through the communication part 909, and / or installed from a removable medium 911. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0158] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 909, and / or installed from the removable medium 911. When the computer program is executed by the processor 901, the above functions defined in the system of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0159] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the client computing device, partially on the client device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the client computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0160] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0161] It will be appreciated by those skilled in the art that the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways, even if such combinations and / or combinations are not explicitly described in the present disclosure. In particular, the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways without departing from the spirit and teachings of the present disclosure. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0162] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. The scope of the present disclosure is defined by the attached claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A method for predicting customer satisfaction, characterized in that: The following steps are involved: Obtain the customer's authorization to obtain personal information; Obtaining personal information of n customers with the authorization of the customer, wherein the personal information of each customer includes N information items, and the N information items are all related to customer satisfaction, where n is an integer greater than or equal to 1, and N is an integer greater than or equal to 2; Quantifying the personal information of the n customers to obtain a personal information matrix, wherein the personal information matrix is ​​a matrix with n rows and N columns, and each row of the personal information matrix represents the quantified personal information of a customer; For the personal information matrix, a graph Laplacian matrix corresponding to the personal information matrix is ​​calculated using a spectral clustering method, wherein the graph Laplacian matrix is ​​a matrix with n rows and n columns; Calculating the spectral density of the graph Laplacian matrix; constructing a graph filter according to the graph Laplacian matrix, wherein, in the process of constructing the graph filter, a passband interval of the graph filter is determined according to the spectral density; Generate a feature matrix according to the graph filter, wherein the feature matrix is ​​a matrix with n rows and b columns, b is a positive integer and 1≤b <n; Based on the feature matrix, classifying the n customers using a clustering method; and Predicting satisfaction of at least one of the n customers according to the classified n customers; The step of calculating the spectral density of the graph Laplacian matrix specifically includes: According to the graph Laplace matrix, calculating the graph Laplace matrix polynomial; calculating a quadratic form of the graph Laplacian matrix polynomial; and Calculating the spectral density of the graph Laplacian matrix according to the quadratic form of the graph Laplacian matrix polynomial; The step of determining the passband interval of the image filter according to the spectral density specifically includes: calculating the eigenvalues ​​of the graph Laplacian matrix according to the spectral density using an iterative method; and The passband interval of the graph filter is determined according to the eigenvalue.

2. The method according to claim 1, characterized in that The adopting of an iterative method to calculate the eigenvalues ​​of the graph Laplacian matrix according to the spectral density specifically includes: Iteratively calculating the upper limit of the interval of the numerical integration of the spectral density according to a specified integration step size; and When the numerical integral of the spectral density is greater than or equal to b, an upper limit of an interval of the numerical integral of the spectral density is determined as the bth eigenvalue of the graph Laplacian matrix.

3. The method according to claim 2, characterized in that The adopting an iterative method to calculate the eigenvalues ​​of the graph Laplacian matrix according to the spectral density further specifically includes: When the iterative calculation is finished, the upper limit of the interval of the numerical integration of the spectral density is determined as the b+1th eigenvalue of the graph Laplacian matrix.

4. The method according to claim 3, characterized in that Determining the passband interval of the graph filter according to the characteristic value specifically includes: An average value of the bth eigenvalue and the b+1th eigenvalue is determined as an upper limit of the passband range of the graph filter.

5. The method according to claim 1, characterized in that The constructing a graph filter according to the graph Laplacian matrix specifically includes: The graph filter is constructed according to the graph Laplacian matrix polynomial and the passband interval of the graph filter.

6. A customer satisfaction prediction device, characterized in that: include: The customer authorization acquisition module is used to obtain the customer's authorization to obtain personal information; A personal information acquisition module, used to: obtain the personal information of n customers with the authorization of the customers to obtain the personal information, wherein the personal information of each customer includes N information items, and the N information items are all related to customer satisfaction, n is an integer greater than or equal to 1, and N is an integer greater than or equal to 2; A personal information matrix acquisition module is used to: quantify the personal information of the n customers to obtain a personal information matrix, wherein the personal information matrix is ​​a matrix of n rows and N columns, and each row of the personal information matrix represents the quantified personal information of a customer; A graph Laplace matrix calculation module, used to: calculate a graph Laplace matrix corresponding to the personal information matrix using a spectral clustering method, wherein the graph Laplace matrix is ​​a matrix with n rows and n columns; A spectral density calculation module, used to calculate the spectral density of the graph Laplacian matrix; A graph filter construction module, configured to: construct a graph filter according to the graph Laplacian matrix, wherein in the process of constructing the graph filter, a passband interval of the graph filter is determined according to the spectral density; A feature matrix acquisition module is used to generate a feature matrix according to the graph filter, wherein the feature matrix is ​​a matrix with n rows and b columns, b is a positive integer and 1≤b <n; A classification module, used to: classify the n customers by using a clustering method based on the feature matrix; and A customer satisfaction prediction module, used to: predict the satisfaction of at least one of the n customers according to the classified n customers; The step of calculating the spectral density of the graph Laplacian matrix specifically includes: According to the graph Laplace matrix, calculating the graph Laplace matrix polynomial; calculating a quadratic form of the graph Laplacian matrix polynomial; and Calculating the spectral density of the graph Laplacian matrix according to the quadratic form of the graph Laplacian matrix polynomial; The step of determining the passband interval of the image filter according to the spectral density specifically includes: calculating the eigenvalues ​​of the graph Laplacian matrix according to the spectral density using an iterative method; and The passband interval of the graph filter is determined according to the eigenvalue.

7. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors execute the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the method according to any one of claims 1 to 5.

9. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images