User portrait construction method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention discloses a user portrait construction method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring user data of a target user from at least two dimensions, and preprocessing the user data; determining a core variable, and constructing an improved customer value model according to the core variable; the core variable comprises a transaction period and the number of types of participated services; determining outlier factors of the target users according to the preprocessed user data, performing clustering analysis on the target users according to the outlier factors of the target users, and determining a clustering result; and analyzing the target users of each cluster according to the improved customer value model and the clustering result, and constructing a user portrait according to the analysis result. By adopting the scheme, the improved customer value model is combined with the outlier factor, the user portrait of the target user is constructed, and more accurate bank business information recommendation is performed for the target user.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of information processing technology, and in particular to a user portrait construction method, device, electronic device and storage medium. Background Art

[0002] At present, my country's large commercial banks have recorded a large amount of customer information data, but lack the organization, induction, cleaning, analysis and utilization of data, lack of data value extraction and personalized service functions. With the continuous development of Internet finance, individual customers have more choices, and banks are in urgent need of proactive and accurate classification marketing for individual customers.

[0003] There are two main types of common bank customer classification methods at present. One type is that bank staff classify customers based on experience or formulate a bank customer classification method by themselves. The other type is a bank customer classification method based on big data mining technology. The bank customer classification system based on big data mining technology uses algorithms and models to analyze bank customer data and then classify bank customers. However, the bank customer classification method classified by bank staff based on experience or formulated by themselves is difficult to sort out the relationship between the complex data and cannot guarantee the accuracy of classification. For the bank customer classification method based on big data mining technology, most of them use big data models and algorithms to analyze customer data, but no reasonable user classification model is formulated before the analysis. The data is only analyzed based on manual classification. Although the accuracy of classification is guaranteed to a certain extent, there is still a lot of room for improvement.

[0004] Therefore, how to accurately classify customers is a technical problem that needs to be urgently solved by technical personnel in this field. Summary of the invention

[0005] The embodiments of the present invention provide a user portrait construction method, device, electronic device and storage medium to improve the accuracy of clustering results and make more accurate banking business information recommendations for target users.

[0006] In a first aspect, an embodiment of the present invention provides a method for constructing a user portrait, comprising:

[0007] Acquire user data of a target user from at least two dimensions, and preprocess the user data;

[0008] Determine core variables and construct an improved customer value model based on the core variables; the core variables include transaction cycle and the number of participating business types;

[0009] Determine the outlier factor of the target user based on the preprocessed user data, perform cluster analysis on the target users based on the outlier factor of each target user and determine the clustering result;

[0010] Analyze the target users of each cluster based on the improved customer value model and clustering results, and build user portraits based on the analysis results.

[0011] In a second aspect, an embodiment of the present invention further provides a user portrait construction device, including:

[0012] A user data processing module, used to obtain user data of a target user from at least two dimensions and pre-process the user data;

[0013] An improved customer value model construction module is used to determine core variables and construct an improved customer value model based on the core variables; the core variables include transaction cycle and the number of participating business types;

[0014] The target user clustering module is used to determine the outlier factor of the target user based on the pre-processed user data, perform cluster analysis on the target users based on the outlier factor of each target user, and determine the clustering result;

[0015] The user portrait construction module is used to analyze the target users of each cluster according to the improved customer value model and the clustering results, and to construct a user portrait according to the analysis results. In a third aspect, an embodiment of the present invention further provides an electronic device, the electronic device comprising:

[0016] one or more processors;

[0017] A storage device for storing one or more programs;

[0018] When the one or more programs are executed by the one or more processors, the one or more processors implement the user portrait construction method described in any embodiment of the present invention.

[0019] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the user portrait construction method described in any embodiment of the present invention.

[0020] In a fifth aspect, an embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the user portrait construction method as described in any embodiment of the present invention.

[0021] The embodiment of the present invention provides a method, device, electronic device and storage medium for constructing a user portrait, which obtains user data of a target user from at least two dimensions and constructs an improved customer value model based on core variables; the core variables include transaction cycle and the number of types of business involved; the outlier factor of the target user is determined based on the preprocessed user data, and the target user is clustered and analyzed based on the outlier factor of each target user to determine the clustering result; the target user of each cluster is analyzed based on the improved customer value model and the clustering result, and a user portrait is constructed based on the analysis result. The technical solution of the embodiment of the present invention is adopted, and labels of different dimensions are marked for the target users of the bank through the label generation method of the improved customer value model; by introducing the concept of outlier factor, the defects of the clustering algorithm are solved to a certain extent, and the accuracy of the clustering result is improved. The improved customer value model is combined with the outlier factor to construct a user portrait of the target user, and more accurate banking business information is recommended for the target user. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts. In the drawings:

[0023] Figure 1 is a flow chart of a method for constructing a user portrait provided in an embodiment of the present invention;

[0024] Figure 2 is a flowchart of another user portrait construction method provided in an embodiment of the present invention;

[0025] Figure 3 is a structural schematic diagram of a user portrait building device provided in an embodiment of the present invention;

[0026] Figure 4 It is a structural schematic diagram of an electronic device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0028] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations (or steps) therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the accompanying drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0029] Among them, the acquisition, storage, use and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations. It should be noted that in the embodiments of this application, some existing solutions in the industry such as certain software, components or models may be mentioned, which should be considered as exemplary, and their purpose is only to illustrate the feasibility of the implementation of the technical solution of this application, but it does not mean that the applicant has or will necessarily use the solution.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention 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 numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein.

[0031] Embodiment 1

[0032] Figure 1 : is a flowchart of a user portrait construction method provided in an embodiment of the present invention. This embodiment is applicable to the case of user portrait construction. The method of this embodiment can be executed by a user portrait construction device, and the device can be implemented in hardware and / or software. The device can be configured in a server for user portrait construction. The method specifically includes the following steps:

[0033] S110: Acquire user data of a target user from at least two dimensions, and preprocess the user data.

[0034] At present, user portraits have been widely studied and applied, and all walks of life have proposed advanced technologies in their own fields for the construction of user portraits. The application practices and research results in the field of medical health mainly focus on how to efficiently process massive medical user data, and on this basis, form patient portraits so as to provide patients with accurate medical services and health information services. The application in the field of e-commerce mainly focuses on the advertising and accurate recommendation of personalized goods or service information based on customer characteristics. The application in the field of tourism is mainly reflected in the accurate recommendation and service of tourism websites and online tourism products and tourist routes, and multi-dimensional and comprehensive tourism user portraits are constructed based on user basic data, interest data and behavior data. The application in the field of libraries mainly focuses on the provision of personalized information recommendation services in libraries, and improves the accuracy of library user information services through user portraits. The embodiment of the present invention intends to take banking business as the background and bank big data as the research basis to study the construction of portraits based on individual bank customers.

[0035] The embodiment of the present invention analyzes the image target based on the bank's massive personal customer data, labels each customer with a specific segmented image, and measures the customer value and the customer's profitability. The embodiment of the present invention obtains the user data of the bank's target users from various dimensions, and the user data characteristics of each dimension are shown in Table 1.

[0036] Table 1 Data characteristics table

[0037]

[0038]

[0039] The acquired user data is preprocessed. In addition to performing operations such as missing value replacement and outlier processing on the user data in the embodiment of the present invention, the user data needs to be normalized to avoid a large data range for the original numerical data. The embodiment of the present invention adopts the max-min normalization method, and its conversion function is as follows:

[0040]

[0041] Where max is the maximum value of the sample data, min is the minimum value of the sample data, x represents the sample data, and x * Represents the standardized sample data.

[0042] S120, determining core variables, and constructing an improved customer value model based on the core variables.

[0043] Among them, the RFM model has been widely studied and applied in terms of customer segmentation and measuring customer value. The RFM model is a widely used marketing model, also known as the customer value model. It groups customers through three indicators: R (Rencency), F (Frequency) and M (Monetary). It is used to measure customer value and revenue-generating ability, use different operating strategies for customers of different values, maximize the company's limited resources, and achieve refined operations. Among them, R (Rencency) represents the time interval between the customer's last purchase of goods, F (Frequency) represents the customer's recent consumption frequency, and M (Monetary) represents the consumption amount in a recent period of time. These three indicators can describe the customer's actual value more objectively through the customer's recent consumption behavior.

[0044] The embodiment of the present invention obtains the core variables of the target user and constructs an improved customer value model based on the core variables; the core variables include the transaction cycle and the number of participating business types. For example, on the basis of the RFM model, two variables, the transaction cycle P (Preriodicity) and the number of participating business types A (Amount), are added to construct an improved customer value model. The time interval R of the customer's most recent transaction, the number of transactions F in the last six months, the transaction amount M in the last six months, the transaction cycle P and the number of participating business types A are selected to form an improved customer value model (RFMPA model).

[0045] In an optional scheme of an embodiment of the present invention, the RFMPA model is only a model label, and the bank's personal customer data needs to be summarized to obtain the specific values ​​of each variable. R, F, and M need to extract the customer ID, transaction time, and transaction amount from the transaction record details table of the bank system; P needs to calculate the transaction time to obtain the standard deviation of the time interval; A needs to extract the business category of each transaction, such as loans, deposits, and product subscriptions. Optionally, after obtaining the specific values ​​of each variable, it is necessary to perform preprocessing operations on the specific values ​​of each variable. The preprocessing operation is consistent with the user data preprocessing operation method of the target user, and is not described in detail in the embodiment of the present invention.

[0046] S130 , determining the outlier factor of the target user according to the preprocessed user data, performing cluster analysis on the target users according to the outlier factor of each target user, and determining a clustering result.

[0047] Among them, the outlier factor is a density-based anomaly detection method used to identify outliers in a data set that are significantly different from other data points, and detects outliers by comparing the local density of each data point with the local density of its neighbors. In the embodiment of the present invention, the outlier factor of the target user is determined, the target user is clustered and the clustering result is determined.

[0048] S140. Analyze the target users of each cluster based on the improved customer value model and the clustering results, and construct a user portrait based on the analysis results.

[0049] Among them, user portrait is the image of the user constructed by data labels. By analyzing massive data information, abstracting the data into labels, and then using these labels to concretize the user image, the user portrait is finally formed. By deeply mining the historical behavioral data of target user characteristics and opinions, a series of labels are created for the user's attributes and behavioral characteristics with the most easy-to-understand vocabulary, and a labeled user role behavioral characteristic model is simply outlined.

[0050] In the embodiment of the present invention, the clustered target users are analyzed through the improved customer value model and the clustering results, and a user portrait is constructed based on the analysis results to push more accurate banking business information to the target users.

[0051] The embodiment of the present invention provides a method for constructing a user portrait, which obtains user data of a target user from at least two dimensions and preprocesses the user data; determines core variables and constructs an improved customer value model based on the core variables; the core variables include transaction cycle and the number of types of participating businesses; determines the outlier factor of the target user based on the preprocessed user data, clusters the target users based on the outlier factors of each target user and determines the clustering results; analyzes the target users of each cluster based on the improved customer value model and the clustering results, and constructs a user portrait based on the analysis results. The technical solution of the embodiment of the present invention is adopted, and labels of different dimensions are added to the target users of the bank through the label generation method of the improved customer value model; by introducing the concept of outlier factors, the defects of the clustering algorithm are solved to a certain extent, and the accuracy of the clustering results is improved. The improved customer value model is combined with the outlier factor to construct a user portrait of the target user, and more accurate banking business information is recommended for the target user.

[0052] Embodiment 2

[0053] Figure 2 Flow chart of a user portrait construction method provided in an embodiment of the present invention. The embodiment of the present invention further optimizes the above embodiment on the basis of the above embodiment, and the embodiment of the present invention can be combined with various optional solutions in one or more of the above embodiments. Figure 2As shown, the user portrait construction method provided in the embodiment of the present invention may include the following steps:

[0054] S210: Acquire user data of a target user from at least two dimensions, and preprocess the user data.

[0055] S220: Determine core variables, and construct an improved customer value model based on the core variables.

[0056] S230: Determine the outlier factor of each target user based on the preprocessed user data.

[0057] Among them, in the K-means algorithm, the K value is the number of clustering categories, which needs to be determined at the beginning. Different initial K values ​​will affect the clustering results. The RFM model divides bank personal customers into 8 categories according to the level of R, F, and M, namely: important value customers, important development customers, important retention customers, important retention customers, general value customers, general development customers, general retention customers, and general retention customers. Among the bank's personal customers, the bank hopes to find important customers and develop potential customers, so the K value is determined to be 3, 4, 5, and 6. Later, the K value with better clustering effect can be selected according to the analysis results.

[0058] After the K value is determined, it is necessary to select the initial cluster center. Generally, there are two methods for selecting the initial cluster center: the synthesis method and the direct selection of actual observation points. However, both methods are prone to misselecting outliers. Therefore, the embodiment of the present invention uses the method of excluding outliers to find the initial cluster center. In order to find outliers, an outlier factor is introduced on the basis of the traditional K-means algorithm. The outlier factor describes the degree of deviation of a data object relative to other data objects. The larger the outlier factor, the higher the degree of deviation of the data point, and the greater the possibility of becoming an outlier. In the embodiment of the present invention, the outlier factor is introduced into the clustering algorithm to solve the defect that the clustering algorithm is prone to misselecting outliers, and improve the accuracy of the clustering results.

[0059] As an optional but non-limiting implementation, the determining of the outlier factor of each target user includes but is not limited to steps A1-A4:

[0060] Step A1: Determine a first distance and a first distance range of a first data object.

[0061] Step A2: Determine the reachable distance and local reachable density between the first data object and the second data object.

[0062] Step A3: Determine a first outlier factor of the first data object according to a first distance range and a local reachability density of the first data object.

[0063] Step A4: Determine a second outlier factor of the second data object according to the method for determining the first outlier factor of the first data object.

[0064] The idea of ​​eliminating outliers and finding the initial cluster center is to first calculate the outlier factor of each data object. In the embodiment of the present invention, the data object refers to the target user, and the target user includes the first data object and the second data object.

[0065] The calculation of the outlier factor of the data object is the most important thing in determining the initial cluster center. The outlier factor is calculated based on the density-based lof algorithm. Taking the first data object p as an example, the outlier factor of the first data object is determined.

[0066] ① kth distance, the kth (positive integer) distance of the first data object p is recorded as dis k (p), for the first data object p and the data object o, if the following two preset conditions are met, then determine the dis k (p) = dis(p, o); wherein dis(p, o) represents the distance between data object p and data object o. The preset condition can be expressed as: A. There are at least k data objects z such that dis(p, z) ≤ dis(p, o); B. There are at least k-1 data objects z such that dis(p, z) < dis(p, o).

[0067] ② k-th distance domain: the set of all data objects contained within the k-th distance of p is defined as the k-th distance domain of p, denoted by N k (p).

[0068] ③ Reachable distance, there are data objects p and q, and the kth distance domain of q is recorded as N k (q), the distance between p and q is denoted as dis(p,q), if p∈N k (q), then the kth reachable distance dis of point p relative to q is r (p, q), the dis r (p, q) is expressed as:

[0069] dis r (p, q) = max{dis k (q), dis(p, q)}

[0070] ④ Local reachability density. For a data object p, the local reachability density lrd(p) of p is calculated as follows:

[0071]

[0072] This formula shows that the local reachability density of a data point p is the inverse of the average reachability density of all objects in the kth domain of point p.

[0073] ⑤ Outlier factor. The outlier factor lof(p) of the data object p is calculated comprehensively based on ② and ④, and the calculation formula is as follows:

[0074]

[0075] Among them, lof(p) reflects the average distribution density of the spatial points within the k-distance range of the data object p. If the value of lof(p) is closer to 1, it indicates that the density of p is more similar to that of its neighboring points, and the possibility of p being an outlier is lower; the smaller the value of lof(p) is than 1, it indicates that the density of p is higher than that of its neighboring points, that is, p is a dense point; then, the larger the value of lof(p) is than 1, the greater the possibility of p being an outlier.

[0076] In summary, using the above calculation formula of the outlier factor, the outlier factor of each data object can be obtained.

[0077] S240. Sort the outlier factors of each target user in ascending order, and select the first target users with the target number to construct an initial clustering center candidate set.

[0078] Among them, after determining the outlier factor of the target user, sort the outlier factors of each target user in ascending order, and select the first target users with the target number to construct an initial clustering center candidate set. For example, select the first a * n (0 < a ≤ 1) data objects to form an initial clustering center candidate set; where n is the size of the sample set, that is, the number of target users.

[0079] S250. Select a preset number of second target users in the initial clustering center candidate set as the initial clustering center, and determine the number of clusters.

[0080] Among them, according to the size of the outlier factor, non-outlier points with smaller outlier factors can be selected as the initial clustering center of the K-means algorithm. In the embodiment of the present invention, a preset number of second target users are selected in the initial clustering center candidate set as the initial clustering center, and the number of clusters is determined; where the number of second target users is the number of clusters.

[0081] As an optional but non-limiting implementation manner, the step of selecting a preset number of second target users in the initial clustering center candidate set as the initial clustering center includes, but is not limited to, steps B1 - B4:

[0082] Step B1: Select any one data object from the initial clustering center candidate set as the third data object.

[0083] Step B2: Select a fourth data object from the initial cluster center candidate set; wherein the fourth data object is farthest from the third data object.

[0084] Step B3: Determine the minimum distance between each data object in the initial cluster center candidate set and the initial center, and select the data object with the largest distance from the minimum distance as the fifth data object.

[0085] Step B4: Determine the sixth data object according to the selection method of the fourth data object; wherein the sum of the number of the third data object, the fourth data object, the fifth data object and the sixth data object is equal to the number of the second target users.

[0086] In which, let C = Φ; randomly select a data object x in F, record it as the third data object C1, and use it as the first cluster center for cluster analysis of data set T. At this time, C = {C1}; C represents the initial cluster center, and F represents the candidate set of initial cluster centers. Find the data point farthest from C1 among the remaining data objects in F, record it as the fourth data object C2, that is, C = {C1, C2}. According to the maximum and minimum principle, find the fifth data object C3 in the candidate set F, that is, C = {C1, C2, C3}. Repeat the selection method of C2 to determine the sixth data object until k initial cluster centers are found.

[0087] Among them, the maximum and minimum principle can be: first calculate x j (x j ∈D / C)(j=1,2,...,n)The minimum distance to the initial center is calculated as:

[0088]

[0089] Then select the data object with the largest distance as C3, and the calculation formula is:

[0090] C i =max{Dis j : j = 1, 2, ..., n}

[0091] Among them, d(C i , x i ) represents the data object C i To the initial center x i Distance j Indicates the minimum distance.

[0092] In order to measure the clustering effect, the embodiment of the present invention refers to the Sum-of-Squared Error Criterion (SSE). The calculation formula of SSE is as follows:

[0093]

[0094] The sse in the above formula calculates the distance from the data object x to its class centroid C i The sum of squared Euclidean distances measures the distance within a class. The smaller the sse value, the smaller the distance within a class, and the better the clustering effect. The larger the K value, the more categories the data is divided into, and the data within the class is relatively more concentrated, that is, the higher the degree of aggregation of each class, and the sum of squared errors will naturally decrease gradually. However, when the K value is smaller, even smaller than the actual number of clusters, the clustering effect is not good, and the corresponding sse will increase. As the K value changes from smaller to larger, the sse value will change from large to small, and will eventually flatten out. This is also the core idea of ​​the elbow method. It is called the "elbow method" because the relationship between sse and K is like an elbow, and the K value corresponding to the elbow is generally the optimal K value.

[0095] S260: cluster the remaining target users in the initial cluster center candidate set using a clustering algorithm, and determine the clustering result.

[0096] The classic K-means clustering algorithm is used to cluster the remaining target users in the initial cluster center candidate set and determine the clustering results.

[0097] As an optional but non-limiting implementation, the clustering algorithm is used to cluster the remaining target users in the initial cluster center candidate set and determine the clustering result, including but not limited to steps C1-C3:

[0098] Step C1: Determine the Euclidean distance between the remaining target users in the initial cluster center candidate set and the initial cluster center, and assign the remaining target users to each cluster based on the minimum distance principle.

[0099] Step C2: Determine the mean of the data in each cluster and use it as the new cluster center.

[0100] Step C3: Repeatedly determine the target user and cluster center in each cluster until the cluster center point remains unchanged, and obtain the target cluster center and clustering results.

[0101] After determining K data objects as the initial cluster centers, the Euclidean distances d(x m , C i ), and the remaining data are assigned to the K clusters closest to the cluster center according to the principle of the most similar data with the smallest distance. Data object x m With cluster center C i The Euclidean distance is as follows:

[0102]

[0103] Calculate the mean of the data in each cluster as the new cluster center The calculation formula is as follows:

[0104]

[0105] Repeat the steps of determining the Euclidean distance and the new cluster center until the condition that the cluster center point remains unchanged is met. At this time, the clustering process is terminated and the clustering result is output.

[0106] S270. Analyze the target users of each cluster based on the improved customer value model and the clustering results, and build a user portrait based on the analysis results.

[0107] Among them, the larger the K value, the more categories the data is divided into, and the data within the class is relatively more concentrated, that is, the higher the degree of aggregation of each class, the error sum of squares will naturally decrease gradually. But when the K value is smaller, even smaller than the actual number of clusters, the clustering effect is not good, and the corresponding SSE will increase. As the K value changes from smaller to larger, the SSE value will change from large to small, and will eventually tend to be flat. This is also the core idea of ​​the elbow method. It is called the "elbow method" because the relationship between SSE and K is like the shape of an elbow, and the K value corresponding to the elbow is generally the optimal K value. In the embodiment of the present invention, taking the K value of 4 as an example, the target users of each cluster are analyzed according to the improved customer value model and the clustering results, and a user portrait is constructed based on the analysis results.

[0108] As an optional but non-limiting implementation, the target users of each cluster are analyzed according to the improved customer value model and the clustering results, and a user portrait is constructed according to the analysis results, including but not limited to steps D1-D2:

[0109] Step D1: Analyze the target users of each cluster according to the improved customer value model and the clustering results, and determine the analysis results of each cluster; the analysis results include customer value characteristics and customer value levels.

[0110] Step D2: construct a user profile based on the analysis results, and recommend corresponding business information to the target user based on the constructed user profile.

[0111] Among them, the user data of the target user is analyzed, and the customer data information includes: the time interval R of the customer's most recent transaction, the number of transactions F in the last six months, the transaction amount M in the last six months, the transaction cycle (i.e., the standard deviation of the transaction time interval in the past year) P, and the number of business types involved in the past year A. After reorganizing and summarizing the clustering results, the user portrait can be constructed, and the clustering results are shown in Table 4.

[0112] Table 4 Clustering results

[0113] Category I Category II Category 3 Category 4 Number of customers 3977 7863 3098 777 Proportion 25％ 50％ 20％ 5％ R 2.098 5.987 10.982 6.176 F 17.11 10.29 12.65 3.87 M 1890982 3908721 2498783 1098723 P 0.019 0.389 0.729 0.398 A 5 3 6 1

[0114] By analyzing the clustering results and characteristic variables in the table, we can get four types of customers as follows:

[0115] The first category is active and valuable customers, accounting for 25% of the total number of customers. The average time interval between recent transactions is 2 days, the transaction frequency and transaction amount are both high, and the standard deviation of transaction time in the past year is relatively small, only 0.019, indicating that the transaction cycle is stable and the number of businesses involved is large. This type of customer should be a VIP customer.

[0116] The second category is non-active important development customers, accounting for 50% of the total number of customers. The average time interval between recent transactions of this type of customer group is 6 days, the transaction amount is high, the number of business types involved is average, the number of transactions in the past six months is relatively small, and the transaction cycle is extremely unstable. This type of customer has great potential and can be developed in the future.

[0117] The third category is active and important customers, accounting for 20% of the total number of customers. This type of customer group has a recent transaction time of more than 10 days, with a high average transaction amount and transaction frequency in the time interval. They participate in a large number of businesses, but their transaction cycle in the past year is extremely unstable. This type of customer should be a loyal old customer who has not conducted banking business for a period of time.

[0118] The fourth category is inactive general customers, accounting for 5% of the total number of customers. The latest transaction interval of this customer group is less than one week, the transaction amount and transaction frequency are in the middle, the number of businesses involved is relatively small, and the transaction cycle is extremely unstable. This type of customer can be regarded as marginal customers or random customers.

[0119] After the user portrait is constructed, the corresponding business information is recommended to the target user based on the constructed user portrait. For example, accurate business push can be performed to the first type of VIP customers and the second type of specific potential customers.

[0120] The embodiment of the present invention provides a method for constructing a user portrait, which obtains user data of a target user from at least two dimensions and preprocesses the user data; determines core variables and constructs an improved customer value model based on the core variables; determines the outlier factor of each target user based on the preprocessed user data; sorts the outlier factor of each target user in ascending order, selects the first target user of the target number to construct an initial cluster center candidate set; selects a preset number of second target users as initial cluster centers in the initial cluster center candidate set, and determines the number of clusters; wherein the number of second target users is the number of clusters; clusters the remaining target users in the initial cluster center candidate set using a clustering algorithm, and determines the clustering results. Analyze the target users of each cluster based on the improved customer value model and the clustering results, and construct a user portrait based on the analysis results. The technical solution of the embodiment of the present invention is adopted to introduce an outlier factor into the traditional clustering algorithm, solve the defect that the clustering algorithm is easy to mistakenly select outliers, and improve the accuracy of the clustering results. Use the improved customer value model and outlier factors to build user profiles of target users and make more accurate banking information recommendations for them.

[0121] Embodiment 3

[0122] Figure 3 : is a schematic diagram of the structure of a user portrait construction device provided in an embodiment of the present invention. The technical solution of this embodiment can be applied to the case of user portrait construction. The device can be implemented by software and / or hardware and is generally integrated on any electronic device with network communication function, including but not limited to: servers, computers, personal digital assistants and other devices. Figure 3 As shown, the user portrait construction device provided in this embodiment may include: a user data processing module 310, an improved customer value model construction module 320, a target user clustering module 330 and a user portrait construction module 340; wherein,

[0123] A user data processing module 310 is used to obtain user data of a target user from at least two dimensions and pre-process the user data;

[0124] An improved customer value model construction module 320 is used to determine core variables and construct an improved customer value model based on the core variables; the core variables include transaction cycle and the number of participating business types;

[0125] The target user clustering module 330 is used to determine the outlier factor of the target user according to the pre-processed user data, perform cluster analysis on the target users according to the outlier factor of each target user and determine the clustering result;

[0126] The user portrait construction module 340 is used to analyze the target users of each cluster according to the improved customer value model and the clustering results, and to construct a user portrait according to the analysis results.

[0127] Based on the above embodiment, optionally, the target user clustering module is specifically used to:

[0128] Determine the outlier factor of each target user based on the preprocessed user data;

[0129] Sort the outlier factors of each target user in ascending order, and select the first target user of the target number to construct the initial cluster center candidate set;

[0130] Selecting a preset number of second target users from the initial cluster center candidate set as initial cluster centers, and determining the number of clusters; wherein the number of second target users is the number of clusters;

[0131] A clustering algorithm is used to cluster the remaining target users in the initial cluster center candidate set and determine the clustering results.

[0132] Based on the above embodiment, optionally, the target user clustering module is further specifically used for:

[0133] determining a first distance and a first distance range of a first data object;

[0134] Determine the reachable distance and local reachable density between the first data object and the second data object:

[0135] Determine a first outlier factor of the first data object according to a first distance field and a local reachable density of the first data object;

[0136] According to the determination method of the first outlier factor of the first data object, the second outlier factor of the second data object is determined; wherein the target user includes the first data object and the second data object, and the data object refers to the target user.

[0137] Based on the above embodiment, optionally, the target user clustering module is further specifically used for:

[0138] Select any data object from the initial cluster center candidate set as the third data object;

[0139] Selecting a fourth data object from the initial cluster center candidate set; wherein the fourth data object is farthest from the third data object;

[0140] Determine the minimum distance between each data object in the initial cluster center candidate set and the initial center, and select the data object with the largest distance from the minimum distance as the fifth data object;

[0141] The sixth data object is determined according to the selection method of the fourth data object; wherein the sum of the numbers of the third data object, the fourth data object, the fifth data object and the sixth data object is equal to the number of the second target users.

[0142] Based on the above embodiment, optionally, the target user clustering module is further specifically used for:

[0143] Determine the Euclidean distance between the remaining target users in the initial cluster center candidate set and the initial cluster center, and assign the remaining target users to each cluster based on the minimum distance principle;

[0144] Determine the mean of the data in each cluster and use it as the new cluster center;

[0145] Repeatedly determine the target users and cluster centers in each cluster until the cluster center point remains unchanged, and obtain the target cluster center and clustering results.

[0146] Based on the above embodiment, optionally, the user portrait construction module is specifically used to:

[0147] Analyze the target users of each cluster according to the improved customer value model and the clustering results, and determine the analysis results of each cluster; the analysis results include customer value characteristics and customer value levels;

[0148] A user profile is constructed based on the analysis results, and corresponding business information is recommended to the target user based on the constructed user profile.

[0149] The user portrait construction device provided in the embodiment of the present invention can execute the user portrait construction method provided in any embodiment of the present invention mentioned above, and has the corresponding functions and beneficial effects of executing the user portrait construction method. For detailed process, please refer to the relevant operations of the user portrait construction method in the above embodiment.

[0150] Embodiment 4

[0151] Figure 4 1 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0152] like Figure 4 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0153] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0154] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the user portrait construction method.

[0155] In some embodiments, the user portrait construction method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the user portrait construction method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the user portrait construction method in any other appropriate manner (e.g., by means of firmware).

[0156] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0157] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0158] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, 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), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0159] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0160] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0161] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0162] Embodiment 5

[0163] An embodiment of the present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the user portrait construction method provided in any embodiment of the present application.

[0164] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0165] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0166] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for constructing a user portrait, characterized in that: The method comprises: Acquire user data of a target user from at least two dimensions, and preprocess the user data; Determine core variables and construct an improved customer value model based on the core variables; the core variables include transaction cycle and the number of participating business types; Determine the outlier factor of the target user based on the preprocessed user data, perform cluster analysis on the target users based on the outlier factor of each target user and determine the clustering result; Analyze the target users of each cluster based on the improved customer value model and clustering results, and build user portraits based on the analysis results.

2. The method according to claim 1, characterized in that The step of determining the outlier factor of the target user based on the preprocessed user data and performing cluster analysis on the target users based on the outlier factor of each target user includes: Determine the outlier factor of each target user based on the preprocessed user data; Sort the outlier factors of each target user in ascending order, and select the first target user of the target number to construct the initial cluster center candidate set; Selecting a preset number of second target users from the initial cluster center candidate set as initial cluster centers, and determining the number of clusters; wherein the number of second target users is the number of clusters; A clustering algorithm is used to cluster the remaining target users in the initial cluster center candidate set and determine the clustering results.

3. The method according to claim 2, characterized in that The determining of the outlier factor of each target user includes: determining a first distance and a first distance range of a first data object; Determine the reachable distance and local reachable density between the first data object and the second data object: Determine a first outlier factor of the first data object according to a first distance field and a local reachable density of the first data object; According to the determination method of the first outlier factor of the first data object, the second outlier factor of the second data object is determined; wherein the target user includes the first data object and the second data object, and the data object refers to the target user.

4. The method according to claim 2, characterized in that: The selecting a preset number of second target users from the initial cluster center candidate set as initial cluster centers includes: Select any data object from the initial cluster center candidate set as the third data object; Selecting a fourth data object from the initial cluster center candidate set; wherein the fourth data object is farthest from the third data object; Determine the minimum distance between each data object in the initial cluster center candidate set and the initial center, and select the data object with the largest distance from the minimum distance as the fifth data object; The sixth data object is determined according to the selection method of the fourth data object; wherein the sum of the numbers of the third data object, the fourth data object, the fifth data object and the sixth data object is equal to the number of the second target users.

5. The method according to claim 2, characterized in that: The clustering algorithm is used to cluster the remaining target users in the initial cluster center candidate set and determine the clustering result, including: Determine the Euclidean distance between the remaining target users in the initial cluster center candidate set and the initial cluster center, and assign the remaining target users to each cluster based on the minimum distance principle; Determine the mean of the data in each cluster and use it as the new cluster center; Repeatedly determine the target users and cluster centers in each cluster until the cluster center point remains unchanged, and obtain the target cluster center and clustering results.

6. The method according to claim 1, characterized in that The target users of each cluster are analyzed according to the improved customer value model and the clustering results, and a user portrait is constructed according to the analysis results, including: Analyze the target users of each cluster according to the improved customer value model and the clustering results, and determine the analysis results of each cluster; the analysis results include customer value characteristics and customer value levels; A user profile is constructed based on the analysis results, and corresponding business information is recommended to the target user based on the constructed user profile.

7. A user portrait construction device, characterized in that: The device comprises: A user data processing module, used to obtain user data of a target user from at least two dimensions and pre-process the user data; An improved customer value model construction module is used to determine core variables and construct an improved customer value model based on the core variables; the core variables include transaction cycle and the number of participating business types; The target user clustering module is used to determine the outlier factor of the target user based on the pre-processed user data, perform cluster analysis on the target users based on the outlier factor of each target user, and determine the clustering result; The user portrait construction module is used to analyze the target users of each cluster based on the improved customer value model and clustering results, and to construct a user portrait based on the analysis results.

8. An electronic device, characterized in that: include: 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 implement the user portrait construction method described in any one of claims 1-6.

9. A storage medium containing computer executable instructions, characterized in that: The computer executable instructions, when executed by a computer processor, are used to execute the user portrait construction method as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for constructing a user portrait according to any one of claims 1 to 6 is implemented.

Images