Company customer group division method, computer program product and electronic device

Abstract

The invention provides a company customer group division method, a computer program product and an electronic device, the company has a plurality of customers, the method comprises: obtaining feature data of each customer, the feature data comprising basic information data and financial information data; performing textualization preprocessing on the feature data to obtain target feature data including a plurality of text features, the text features including feature names and feature values; performing vectorization processing on the target feature data to obtain a feature vector; determining one of the plurality of customers as a reference target customer, and calculating a weighted similarity score of each customer and the reference target customer at least according to the feature vector; and according to each weighted similarity score, screening out a target number of customers from the plurality of customers to obtain a target customer group so as to provide corresponding financial services for the target customer group. According to the method and the device, the problem that the accuracy of financial services provided for different customer groups is relatively low due to relatively low correlation and accuracy of customer group division of a bank company in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the technical field of customer group segmentation, and in particular, to a company customer group segmentation method, a computer-readable storage medium, a computer program product, and an electronic device. Background Art

[0002] The current banking industry is increasingly turning to the personalization and customization of "customer groups and customers". The traditional business strategy of facing all customers is no longer suitable for the current development needs of banks. Especially for corporate customers, in order to reduce the risk of losing a single large customer, gradually introduce customers such as commercial companies or individual industrial and commercial households of various types and development stages. For a variety of customers with different development scales, industries and profit expectations, it is necessary to provide differentiated financial services to customer groups and carry out targeted activities for customers in the same customer group.

[0003] At present, customer segmentation mainly adopts the method of manually establishing limited thresholds for numerical features, and manually labeling the delineation results. This customer segmentation method is relatively simple, ignoring the rich text information of corporate customers, such as company profiles, annual report information, company names, etc. The semantic features of text information are ignored. For example, "** Garment Factory" and "** Clothing Factory" may be divided into different customer groups due to artificial thresholds, and the correlation between the two is not reflected. The relevance and accuracy of the target customer selection results using the current mainstream method are not satisfactory, which restricts the accuracy of providing financial services. Summary of the invention

[0004] The main purpose of the present application is to provide a corporate customer segmentation method, a computer-readable storage medium, a computer program product and an electronic device, so as to at least solve the problem in the prior art that the relevance and accuracy of bank corporate customer segmentation are low, resulting in low accuracy of financial services provided to different customer groups.

[0005] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for dividing a company's customer groups is provided, wherein the company has multiple customers, and the method comprises: obtaining feature data of each of the customers, wherein the feature data comprises basic information data and financial information data, wherein the basic information data is data representing the attribute information of the customer, and the financial information data is data representing the financial status and economic activities of the customer; performing textual preprocessing on the feature data to obtain target feature data comprising multiple text features, wherein the text features comprise feature names and feature values, wherein the feature names represent identifiers of a certain attribute of the customer, and the feature values ​​represent the information of the customer under the corresponding feature names; performing vectorization processing on the target feature data to obtain a feature vector; determining one of the multiple customers as a reference target customer, and calculating a weighted similarity score between each of the customers and the reference target customer based at least on the feature vector; and screening out a target number of customers from the multiple customers based on each of the weighted similarity scores to obtain a target customer group, so as to provide corresponding financial services to the target customer group.

[0006] Optionally, vectorizing the target feature data to obtain a feature vector includes: training the target feature data using a Doc2Vec model to obtain the feature vector.

[0007] Optionally, the feature data is preprocessed into text form to obtain target feature data including multiple text features, including: preprocessing the feature data into text form to obtain multiple text features; and concatenating the multiple text features corresponding to the customer into paragraphs to obtain the target feature data.

[0008] Optionally, after acquiring the characteristic data of each of the customers and before performing textual preprocessing on the characteristic data, the method further includes: performing data cleaning on the characteristic data to remove invalid characteristic data in the characteristic data, and performing textual preprocessing on the characteristic data, including: performing textual preprocessing on the characteristic data after data cleaning.

[0009] Optionally, the text features are divided into numerical features and text features, the numerical features represent features that can be quantified, and the text features represent features that exist in text form, and at least based on the feature vector, the weighted similarity score between each of the customers and the reference target customer is calculated, including: using the weighted similarity formula Calculate the weighted similarity score between the customer and the reference target customer, where Score represents the weighted similarity score, A represents the feature vector corresponding to the customer, I represents the feature vector corresponding to the reference target customer, b1, b2, ..., bm represents m weighted numerical features corresponding to the customer, wherein the weighted numerical features are obtained after weighting the numerical features.

[0010] Optionally, after the feature data is preprocessed in text form, before the weighted similarity score between the customer and the reference target customer is calculated, the method further comprises: using formula b m =β m *α m , calculate the weighted numerical features, where β m ∈(0,1],β m Indicates the weight value corresponding to the numerical feature, α m are the normalized eigenvalues ​​of the m numerical features.

[0011] Optionally, based on each of the weighted similarity scores, a target number of the customers are screened out from the multiple customers to obtain a target customer group, including: sorting and numbering the multiple weighted similarity scores in order from large to small to obtain a score sequence; determining that the customers corresponding to the first n weighted similarity scores in the score sequence are target customers, wherein n target customers constitute the target customer group, and n represents the target number.

[0012] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute any one of the company customer segmentation methods.

[0013] According to another aspect of the present application, a computer program product is provided, comprising computer instructions, wherein when the computer instructions are executed by a processor, any one of the methods for segmenting company customer groups is implemented.

[0014] According to another aspect of the present application, an electronic device is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include means for executing any one of the company customer segmentation methods.

[0015] Applying the technical solution of the present application, a company has multiple customers. First, feature data including basic information data and financial information data of each customer is obtained, and then the feature data is pre-processed into text to obtain target feature data including multiple text features, where the text features include feature names and feature values. The target feature data is then vectorized to obtain a feature vector, and one of the multiple customers is determined as a reference target customer. At least based on the feature vector, a weighted similarity score between each customer and the reference target customer is calculated. Finally, based on each weighted similarity score, a target number of customers are screened out from the multiple customers to obtain a target customer group, so as to provide corresponding financial services to the target customer group. Compared with the problem in the prior art that the relevance and accuracy of bank company customer group division are low, resulting in low accuracy of financial services provided to different customer groups, the present application performs textual preprocessing on feature data to obtain target feature data including multiple text features, and the text features include feature names and feature values. That is, by combining the feature names and feature values ​​in a textual preprocessing manner, the company's customer text information and semantic information are fully utilized to improve the feature richness. In addition, the weighted similarity score indicator is used to screen the target customer group, so that the weighted calculation of the focus feature values ​​can be combined with the specific business scenarios, thereby ensuring that the screened target customer group is relatively accurate, and thus ensuring that the financial services provided to the target customer group are highly accurate and that customer satisfaction is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings constituting part of the present application are used to provide a further understanding of the present application. The exemplary embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0017] Figure 1 A hardware structure block diagram of a mobile terminal for executing a method for dividing company customer groups provided in an embodiment of the present application is shown;

[0018] Figure 2 A flow chart of a method for dividing company customer groups provided according to an embodiment of the present application is shown.

[0019] The above drawings include the following reference numerals:

[0020] 102, processor; 104, memory; 106, transmission device; 108, input and output devices. DETAILED DESCRIPTION

[0021] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0022] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] As introduced in the background technology, the prior art has low relevance and accuracy in the segmentation of bank corporate customer groups, resulting in low precision of financial services provided to different customer groups. To solve the above problems, the embodiments of the present application provide a corporate customer segmentation method, computer-readable storage medium, computer program product and electronic device.

[0025] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0026] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 FIG. 1 is a hardware structure block diagram of a mobile terminal of a method for dividing company customer groups according to an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1Different configurations shown.

[0027] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the company customer segmentation method in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, the above method is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above-mentioned network include but are not limited to the Internet, corporate intranets, local area networks, mobile communication networks and combinations thereof. The transmission device 106 is used to receive or send data via a network. The above-mentioned specific examples of the network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0028] In this embodiment, a method for segmenting company customers that runs on a mobile terminal, a computer terminal, or a similar computing device is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0029] Figure 2 is a flow chart of a method for dividing company customer groups according to an embodiment of the present application. A company has multiple customers, such as Figure 2 As shown, the method comprises the following steps:

[0030] Step S201, acquiring characteristic data of each of the above customers, the above characteristic data including basic information data and financial information data, the above basic information data is data representing attribute information of the above customers, and the above financial information data is data representing financial status and economic activities of the above customers;

[0031] Step S202, performing text preprocessing on the feature data to obtain target feature data including multiple text features, wherein the text features include feature names and feature values, wherein the feature names represent identifiers of a certain attribute of the customer, and the feature values ​​represent information of the customer under the corresponding feature names;

[0032] Step S203, performing vectorization processing on the target feature data to obtain a feature vector;

[0033] Step S204, determining one of the multiple customers as a reference target customer, and calculating a weighted similarity score between each of the customers and the reference target customer based at least on the feature vector;

[0034] Step S205 , screening a target number of the above-mentioned customers from the plurality of the above-mentioned customers according to the weighted similarity scores to obtain a target customer group, so as to provide corresponding financial services to the above-mentioned target customer group.

[0035] Through the above embodiment, the company has multiple customers. First, feature data including basic information data and financial information data of each customer is obtained, and then the feature data is pre-processed into text to obtain target feature data including multiple text features, and the text features include feature names and feature values. The target feature data is then vectorized to obtain a feature vector, and one of the multiple customers is determined as a reference target customer. At least based on the feature vector, a weighted similarity score between each customer and the reference target customer is calculated. Finally, based on each weighted similarity score, a target number of customers are screened out from the multiple customers to obtain a target customer group, so as to provide corresponding financial services to the target customer group. Compared with the problem in the prior art that the relevance and accuracy of bank company customer group division are low, resulting in low accuracy of financial services provided to different customer groups, the present application performs textual preprocessing on feature data to obtain target feature data including multiple text features, and the text features include feature names and feature values. That is, by combining the feature names and feature values ​​in a textual preprocessing manner, the company's customer text information and semantic information are fully utilized to improve the feature richness. In addition, the weighted similarity score indicator is used to screen the target customer group, so that the weighted calculation of the focus feature values ​​can be combined with the specific business scenarios, thereby ensuring that the screened target customer group is relatively accurate, and thus ensuring that the financial services provided to the target customer group are highly accurate and that customer satisfaction is high.

[0036] Specifically, the above-mentioned basic information data refers to data on customer attributes such as customer name, certificate, registered address, legal person information and business scope; the above-mentioned financial information data refers to financial data reflected in financial reports such as the company charter, annual income, annual profit, growth rate, profit margin and debt ratio in the annual and quarterly reports of the customer company.

[0037] Specifically, the above feature names refer to specific identifiers that can describe a certain attribute of a customer. These names are classification labels for customer data, used to identify a column or attribute in the data set, such as "province", "industry category", "company name", etc. The above feature names are a classification of customer information, which tells us what each field in the data set represents. The above feature values ​​refer to the specific data values ​​corresponding to the feature names. These values ​​are actual observation data, which represent the specific information of the customer under a certain feature name. For example, if the feature name is "province", the feature value may be a specific geographical location name such as "Beijing" or "Shanghai". If the feature name is "industry category", the feature value may be an industry classification name such as "manufacturing" or "financial services". Specifically, for example: if the feature name is "province", the corresponding feature value is "Beijing", and the text feature is "province Beijing".

[0038] Specifically, the target number is less than the total number of customers. In actual application, those skilled in the art may set the target number based on experience or obtain it through multiple experiments, and this application does not impose any specific restrictions on this.

[0039] In an optional solution, the target feature data is vectorized to obtain a feature vector, including: using a Doc2Vec model to train the target feature data to obtain the feature vector. In this embodiment, the Doc2Vec model can convert the target feature data into numerical vectors, which can capture the semantic information in the text. In this way, the unstructured target feature data can be converted into structured data that can be processed by a machine learning algorithm, preparing for subsequent weighted similarity calculations.

[0040] Specifically, the Doc2Vec model is trained to obtain a vectorized representation of the target feature data (i.e., a feature vector), that is, the target feature data of each customer is represented by a set of feature vectors.

[0041] Specifically, the precision of model parameters is proportional to the training data (i.e., target feature data). The more training data there is, the better the training effect will be.

[0042] It should be noted that the present application may also use other machine learning models to perform vector processing on the target feature data, such as the BERT model and the Word2Vec model, etc., and the present application does not impose any specific restrictions on this.

[0043] In other embodiments, the feature data is preprocessed into text to obtain target feature data including multiple text features, including: preprocessing the feature data into text to obtain multiple text features; splicing the multiple text features corresponding to the customer into paragraphs to obtain the target feature data. In this embodiment, splicing multiple text features of the customer into paragraphs helps to construct a comprehensive customer feature description, which can more comprehensively reflect the characteristics and needs of the customer, thereby further ensuring that the target customer group screened out later is relatively accurate.

[0044] Specifically, multiple text features corresponding to a single customer are concatenated as a paragraph to obtain target feature data corresponding to the single customer.

[0045] In some other optional solutions, after acquiring the characteristic data of each of the above-mentioned customers, before performing textual preprocessing on the above-mentioned characteristic data, the above-mentioned method further includes: performing data cleaning on the above-mentioned characteristic data to remove invalid characteristic data in the above-mentioned characteristic data, and performing textual preprocessing on the above-mentioned characteristic data, including: performing textual preprocessing on the above-mentioned characteristic data after data cleaning. In this embodiment, by removing invalid characteristic data through data cleaning, it can be ensured that the data processed subsequently is accurate and usable, thereby improving the quality and reliability of data analysis, and further ensuring that the target customer group screened subsequently is relatively accurate.

[0046] In some further optional solutions, the text features are divided into numerical features and text features, the numerical features represent features that can be quantified, and the text features represent features that exist in text form, and at least based on the feature vector, the weighted similarity score between each of the above customers and the above reference target customer is calculated, including: using the weighted similarity formula Calculate the weighted similarity score between the customer and the reference target customer, where Score represents the weighted similarity score, A represents the feature vector corresponding to the customer, I represents the feature vector corresponding to the reference target customer, b1, b2, ..., b m represents the m weighted numerical features corresponding to the above-mentioned customer, and the above-mentioned weighted numerical features are obtained after weighting the above-mentioned numerical features. In this embodiment, a specific weighted similarity formula is used to calculate the weighted similarity score between the customer and the reference target customer, which ensures that the obtained weighted similarity score is relatively accurate, thereby further ensuring that the target customer group screened out subsequently is relatively accurate.

[0047] In some exemplary embodiments, after the feature data is preprocessed into text form, before the weighted similarity score between the customer and the reference target customer is calculated, the method further includes: using formula b m =β m *αm , calculate the above weighted numerical features, where β m ∈(0,1],β m Represents the weight value corresponding to the above numerical features, α m is the normalized feature value of the m numerical features. In this embodiment, a specific formula is used to calculate the weighted numerical features, in preparation for the subsequent calculation of the weighted similarity score.

[0048] Specifically, β m is the experience weight, β m The default value is 1.

[0049] Specifically, numerical features are usually expressed in the form of numbers and can be processed through mathematical operations. In the field of customer analysis of banking companies, numerical features include but are not limited to customers' income, age, asset size, revenue, net profit, etc. Text features usually contain rich semantic information and need to be extracted and analyzed through text processing technology. Text features may include but are not limited to customers' company names, company profiles, annual report information, etc. These features are unstructured and need to be vectorized through specific algorithms (such as Doc2Vec models) to convert text information into numerical forms that can be mathematically operated. The vectorized representation of text features makes it possible to use semantic similarity in customer segmentation, improving the relevance and accuracy of customer segmentation.

[0050] Specifically, in the present invention, by combining feature names and feature values ​​into text features and using Doc2Vec model training to obtain vectorized representation of target feature data, full utilization of company customer text information and semantic information can be achieved. This method not only takes into account numerical features, but also combines text features, thereby improving the richness of features and the accuracy of customer segmentation.

[0051] Specifically, for indicators that the business is more concerned about, such as enterprise scale, enterprise revenue, enterprise net profit and other numerical characteristics, β can be increased according to the emphasis m The weight is then increased to highlight the weight of the feature in the similarity calculation, and the selected target customer group is more suitable for specific business scenarios. For example, if the annual net profit indicator of the enterprise is considered, the weight of the annual net profit can be increased to twice the initial value.

[0052] In some other exemplary embodiments, according to each of the weighted similarity scores, a target number of the above-mentioned customers are screened out from the plurality of the above-mentioned customers to obtain a target customer group, including: sorting and numbering the plurality of the above-mentioned weighted similarity scores in order from large to small to obtain a score sequence; determining the above-mentioned customers corresponding to the first n weighted similarity scores in the above-mentioned score sequence as target customers, wherein n of the above-mentioned target customers constitute the above-mentioned target customer group, and n represents the above-mentioned target number. In this embodiment, by calculating the weighted similarity score of the customer feature data and sorting the customers according to the score, the customer group with the highest similarity to the reference target customer can be screened out more accurately, thereby achieving a more accurate division of customer groups.

[0053] Specifically, according to the sorting of weighted similarity scores, customers with high similarity to the reference target customers are screened out to obtain the target customer group. Through the representation of text features, semantic representation is added to the similarity calculation, for example, the similarity scores of customers with similar semantics such as clothing factories and garment factories are improved.

[0054] Specifically, the present invention makes full use of the textual features of the company's customers, such as company name, profile, and financial report data, to characterize the customers with feature vectors. It not only uses pure data features, but also uses semantic analysis as customer features as model input, thereby improving feature richness. In addition, the target customer group can be screened out by using the weighted similarity score index. The weighted calculation of the feature values ​​can be combined with specific business scenarios, and the screening results can be more inclined towards the target customer group.

[0055] In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the company customer group segmentation method of the present application will be described in detail below in conjunction with specific embodiments.

[0056] This embodiment relates to a specific method for dividing company customer groups. The company has multiple customers, and specifically includes the following steps:

[0057] Step S1: Acquire characteristic data of each customer, the characteristic data including basic information data and financial information data, the basic information data is data representing the attribute information of the customer, and the financial information data is data representing the financial status and economic activities of the customer;

[0058] Step S2: performing data cleaning on the feature data to remove invalid feature data in the feature data;

[0059] Step S3: Perform text preprocessing on the cleaned feature data to obtain multiple text features, and concatenate the multiple text features corresponding to the customer as paragraphs to obtain target feature data. The text features include feature names and feature values. The feature name represents an identifier of a certain attribute of the customer, and the feature value represents the information of the customer under the corresponding feature name.

[0060] Step S4: Use the Doc2Vec model to train the target feature data to obtain a feature vector;

[0061] Step S5: determining one of the multiple customers as a reference target customer, and calculating a weighted similarity score between each customer and the reference target customer based at least on the feature vector;

[0062] Step S6: According to each weighted similarity score, a target number of customers are screened out from multiple customers to obtain a target customer group, so as to provide corresponding financial services to the target customer group.

[0063] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0064] An embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the company customer segmentation method.

[0065] Specifically, the company's customer segmentation methods include:

[0066] Step S201, acquiring characteristic data of each of the above customers, the above characteristic data including basic information data and financial information data, the above basic information data is data representing attribute information of the above customers, and the above financial information data is data representing financial status and economic activities of the above customers;

[0067] Step S202, performing text preprocessing on the feature data to obtain target feature data including multiple text features, wherein the text features include feature names and feature values, wherein the feature names represent identifiers of a certain attribute of the customer, and the feature values ​​represent information of the customer under the corresponding feature names;

[0068] Step S203, performing vectorization processing on the target feature data to obtain a feature vector;

[0069] Step S204, determining one of the multiple customers as a reference target customer, and calculating a weighted similarity score between each of the customers and the reference target customer based at least on the feature vector;

[0070] Step S205 , screening a target number of the above-mentioned customers from the plurality of the above-mentioned customers according to the weighted similarity scores to obtain a target customer group, so as to provide corresponding financial services to the above-mentioned target customer group.

[0071] Optionally, vectorizing the target feature data to obtain a feature vector includes: training the target feature data using a Doc2Vec model to obtain the feature vector.

[0072] Optionally, the feature data is preprocessed into text form to obtain target feature data including multiple text features, including: preprocessing the feature data into text form to obtain multiple text features; concatenating the multiple text features corresponding to the customer into paragraphs to obtain the target feature data.

[0073] Optionally, after obtaining the characteristic data of each of the above-mentioned customers and before performing textual preprocessing on the above-mentioned characteristic data, the above-mentioned method also includes: performing data cleaning on the above-mentioned characteristic data to remove invalid characteristic data in the above-mentioned characteristic data, and performing textual preprocessing on the above-mentioned characteristic data, including: performing textual preprocessing on the above-mentioned characteristic data after data cleaning.

[0074] Optionally, the text features are divided into numerical features and text features, the numerical features represent features that can be quantified, and the text features represent features that exist in text form, and at least based on the feature vector, the weighted similarity score between each of the customers and the reference target customer is calculated, including: using the weighted similarity formula Calculate the weighted similarity score between the customer and the reference target customer, where Score represents the weighted similarity score, A represents the feature vector corresponding to the customer, I represents the feature vector corresponding to the reference target customer, b1, b2, ..., b m It represents m weighted numerical features corresponding to the above customers, and the above weighted numerical features are obtained after weighted processing of the above numerical features.

[0075] Optionally, after the feature data is preprocessed into text form, before calculating the weighted similarity score between the customer and the reference target customer, the method further includes: using formula b m =β m *α m , calculate the above weighted numerical features, where β m ∈(0,1],β m Represents the weight value corresponding to the above numerical features, α m is the normalized eigenvalue of the m numerical features mentioned above.

[0076] Optionally, based on each of the weighted similarity scores, a target number of customers are screened out from the multiple customers to obtain a target customer group, including: sorting and numbering the multiple weighted similarity scores in order from large to small to obtain a score sequence; determining the customers corresponding to the first n weighted similarity scores in the score sequence as target customers, wherein n target customers constitute the target customer group, and n represents the target number.

[0077] The present application also provides a computer program product, including computer instructions, which, when executed by a processor, implement at least the following method steps: step S201, obtaining feature data of each of the above-mentioned customers, the above-mentioned feature data including basic information data and financial information data, the above-mentioned basic information data is data representing the attribute information of the above-mentioned customers, and the above-mentioned financial information data is data representing the financial status and economic activities of the above-mentioned customers; step S202, performing text preprocessing on the above-mentioned feature data to obtain target feature data including multiple text features, the above-mentioned text features include feature names and feature values, the above-mentioned feature names represent identifiers of a certain attribute of the above-mentioned customers, and the above-mentioned feature values ​​represent the information of the above-mentioned customers under the corresponding above-mentioned feature names; step S203, performing vectorization processing on the above-mentioned target feature data to obtain a feature vector; step S204, determining that one of the above-mentioned customers among the multiple above-mentioned customers is a reference target customer, and at least based on the above-mentioned feature vector, calculating the weighted similarity score between each of the above-mentioned customers and the above-mentioned reference target customer; step S205, based on each of the above-mentioned weighted similarity scores, screening out a target number of the above-mentioned customers from the multiple above-mentioned customers to obtain a target customer group, so as to provide corresponding financial services to the above-mentioned target customer group.

[0078] Optionally, vectorizing the target feature data to obtain a feature vector includes: training the target feature data using a Doc2Vec model to obtain the feature vector.

[0079] Optionally, the feature data is preprocessed into text form to obtain target feature data including multiple text features, including: preprocessing the feature data into text form to obtain multiple text features; concatenating the multiple text features corresponding to the customer into paragraphs to obtain the target feature data.

[0080] Optionally, after obtaining the characteristic data of each of the above-mentioned customers and before performing textual preprocessing on the above-mentioned characteristic data, the above-mentioned method also includes: performing data cleaning on the above-mentioned characteristic data to remove invalid characteristic data in the above-mentioned characteristic data, and performing textual preprocessing on the above-mentioned characteristic data, including: performing textual preprocessing on the above-mentioned characteristic data after data cleaning.

[0081] Optionally, the text features are divided into numerical features and text features, the numerical features represent features that can be quantified, and the text features represent features that exist in text form, and at least based on the feature vector, the weighted similarity score between each of the customers and the reference target customer is calculated, including: using the weighted similarity formula Calculate the weighted similarity score between the customer and the reference target customer, where Score represents the weighted similarity score, A represents the feature vector corresponding to the customer, I represents the feature vector corresponding to the reference target customer, b1, b2, ..., b m It represents m weighted numerical features corresponding to the above customers, and the above weighted numerical features are obtained after weighted processing of the above numerical features.

[0082] Optionally, after the feature data is preprocessed into text form, before calculating the weighted similarity score between the customer and the reference target customer, the method further includes: using formula b m =β m *α m , calculate the above weighted numerical features, where β m ∈(0,1],β m Represents the weight value corresponding to the above numerical features, α m is the normalized eigenvalue of the m numerical features mentioned above.

[0083] Optionally, based on each of the weighted similarity scores, a target number of customers are screened out from the multiple customers to obtain a target customer group, including: sorting and numbering the multiple weighted similarity scores in order from large to small to obtain a score sequence; determining the customers corresponding to the first n weighted similarity scores in the score sequence as target customers, wherein n target customers constitute the target customer group, and n represents the target number.

[0084] The present application also provides an electronic device, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include means for executing any one of the above-mentioned company customer segmentation methods.

[0085] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0086] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0087] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0088] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0089] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0090] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0091] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0092] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0093] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0094] From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:

[0095] In the company customer segmentation method of the present application, the company has multiple customers. First, feature data including basic information data and financial information data of each customer is obtained, and then the feature data is pre-processed into text to obtain target feature data including multiple text features, where the text features include feature names and feature values. The target feature data is then vectorized to obtain a feature vector, and one of the multiple customers is determined as a reference target customer. At least based on the feature vector, a weighted similarity score between each customer and the reference target customer is calculated. Finally, based on each weighted similarity score, a target number of customers are screened out from the multiple customers to obtain a target customer group, so as to provide corresponding financial services to the target customer group. Compared with the problem in the prior art that the relevance and accuracy of bank company customer group division are low, resulting in low accuracy of financial services provided to different customer groups, the present application performs textual preprocessing on feature data to obtain target feature data including multiple text features, and the text features include feature names and feature values. That is, by combining the feature names and feature values ​​in a textual preprocessing manner, the company's customer text information and semantic information are fully utilized to improve the feature richness. In addition, the weighted similarity score indicator is used to screen the target customer group, so that the weighted calculation of the focus feature values ​​can be combined with the specific business scenarios, thereby ensuring that the screened target customer group is relatively accurate, and thus ensuring that the financial services provided to the target customer group are highly accurate and that customer satisfaction is high.

[0096] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for dividing company customer groups, characterized in that: The company has multiple customers, and the method includes: Acquire characteristic data of each of the customers, the characteristic data including basic information data and financial information data, the basic information data is data representing attribute information of the customer, and the financial information data is data representing financial status and economic activities of the customer; Performing text preprocessing on the feature data to obtain target feature data including a plurality of text features, wherein the text features include a feature name and a feature value, wherein the feature name represents an identifier of a certain attribute of the customer, and the feature value represents information of the customer under the corresponding feature name; Performing vectorization processing on the target feature data to obtain a feature vector; Determine one of the multiple customers as a reference target customer, and calculate a weighted similarity score between each of the customers and the reference target customer based at least on the feature vector; According to each of the weighted similarity scores, a target number of the customers are screened out from the plurality of customers to obtain a target customer group, so as to provide corresponding financial services to the target customer group.

2. The method for dividing company customer groups according to claim 1, characterized in that: The target feature data is vectorized to obtain a feature vector, including: The target feature data is trained using the Doc2Vec model to obtain the feature vector.

3. The method for dividing company customer groups according to claim 1, characterized in that: The feature data is preprocessed into text to obtain target feature data including multiple text features, including: Performing text preprocessing on the feature data to obtain a plurality of text features; The plurality of text features corresponding to the customer are concatenated as a paragraph to obtain the target feature data.

4. The method for dividing company customer groups according to claim 1, characterized in that: After acquiring the characteristic data of each customer and before performing textual preprocessing on the characteristic data, the method further includes: performing data cleaning on the characteristic data to remove invalid characteristic data in the characteristic data, The feature data is preprocessed into text, including: the feature data after data cleaning is preprocessed into text.

5. The method for dividing company customer groups according to claim 1, characterized in that: The text features are divided into numerical features and text features, the numerical features represent features that can be quantified, and the text features represent features that exist in text form. At least based on the feature vector, a weighted similarity score between each of the customers and the reference target customer is calculated, including: Using the weighted similarity formula Calculate the weighted similarity score between the customer and the reference target customer, where Score represents the weighted similarity score, A represents the feature vector corresponding to the customer, I represents the feature vector corresponding to the reference target customer, b1, b2, ..., b m represents m weighted numerical features corresponding to the customer, wherein the weighted numerical features are obtained after weighting the numerical features.

6. The method for dividing company customer groups according to claim 5, characterized in that: After textual preprocessing of the feature data, and before calculating the weighted similarity score between the customer and the reference target customer, the method further includes: Using formula b m =β m *α m , calculate the weighted numerical features, where β m ∈(0,1],β m Indicates the weight value corresponding to the numerical feature, α m are the normalized eigenvalues ​​of the m numerical features.

7. The method for dividing company customer groups according to claim 1, characterized in that: According to each of the weighted similarity scores, a target number of the customers are screened out from the plurality of customers to obtain a target customer group, including: Sorting and numbering the plurality of weighted similarity scores in descending order to obtain a score sequence; The customers corresponding to the first n weighted similarity scores in the score sequence are determined as target customers, wherein n target customers constitute the target customer group, and n represents the target number.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the company customer segmentation method according to any one of claims 1 to 7.

9. A computer program product comprising computer instructions, characterized in that When the computer instructions are executed by a processor, the method for segmenting company customers as described in any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include methods for executing the company customer segmentation method described in any one of claims 1 to 7.

Images