Customer channel matching method, device, computer equipment and storage medium

Abstract

The present invention relates to the field of artificial intelligence and financial technology, and discloses a customer channel matching method, device, computer equipment and storage medium. The method generates comprehensive portrait features of the customer to be matched by acquiring customer portrait data and customer channel behavior data of the customer to be matched, and inputs the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain recommended marketing channels for the customer to be matched, so as to match the customer with a suitable marketing channel. By selecting a suitable marketing channel for the customer to be matched based on the unique customer data of different customers through a pre-trained preset matching model, the customer can be given a more matching marketing channel, thereby increasing the possibility of the customer receiving more interesting information, avoiding the customer's aversion to the content of the material sent by the channel, improving the customer's operation results, and facilitating the improvement of the customer's substantial conversion.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and financial technology, and in particular to a customer channel matching method, device, computer equipment and storage medium. Background Art

[0002] At present, the relationship between bank retail business and customers is becoming increasingly close, and it can involve many solutions in customers' lives, including but not limited to investment and financial management, insurance purchase, mobile payment, etc. Limited by the large customer base targeted by bank retail business, it is difficult to manually specify a unique customer marketing method for each customer. In traditional customer marketing plans, due to the limited human resources of banks and the large number of customers, it is usually difficult to consider the feelings and preferences of all customers. In order to save costs and quickly reach them, it is usually necessary to blindly select customer groups according to business rules and reach them through the same popular channel content.

[0003] The inventors realized that traditional customer marketing programs often result in some customers not receiving information of interest, or even being disgusted with the content of materials sent by the channel, resulting in a waste of operating resources. Summary of the invention

[0004] The present invention provides a customer channel matching method, device, computer equipment and storage medium to solve the technical problem that traditional customer marketing solutions blindly select customer groups according to business rules and conduct stereotyped popular channel content reach, resulting in some customers not receiving information of interest or even being disgusted with the material content sent by the channel, resulting in a waste of operating resources.

[0005] In a first aspect, a customer channel matching method is provided, comprising: obtaining customer portrait data to be matched and customer channel behavior data to be matched of the customer to be matched; generating comprehensive portrait features of the customer to be matched based on the customer portrait data to be matched and the customer channel behavior data to be matched; inputting the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain recommended marketing channels for the customer to be matched, so as to serve as suitable marketing channels for matching the customer to be matched, wherein the pre-trained preset matching model is obtained by training an improved boosting decision tree algorithm model by constructing a positive and negative sample sequence matrix through historical conversion data of multiple sample customers in multiple marketing channels.

[0006] In a second aspect, a customer channel matching device is provided, including: an acquisition module for acquiring customer portrait data to be matched and customer channel behavior data to be matched of the customer to be matched; a generation module for generating comprehensive portrait features of the customer to be matched based on the customer portrait data to be matched and the customer channel behavior data to be matched; a matching module for inputting the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain recommended marketing channels for the customer to be matched, so as to serve as suitable marketing channels for matching the customer to be matched, wherein the pre-trained preset matching model is obtained by training an improved boosting decision tree algorithm model by constructing a positive and negative sample sequence matrix based on historical conversion data of multiple sample customers in multiple marketing channels.

[0007] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned customer channel matching method when executing the computer program.

[0008] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned customer channel matching method are implemented.

[0009] In the scheme implemented by the above-mentioned customer channel matching method, device, computer equipment and storage medium, the method can generate comprehensive portrait features of the customer to be matched by obtaining the customer portrait data and channel behavior data of the customer to be matched, and input the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain recommended marketing channels for the customer to be matched, so as to match the appropriate marketing channels for the customer to be matched. By selecting appropriate marketing channels based on the unique customer data of different customers through a pre-trained preset matching model, the customer can be given a more matching marketing channel, thereby increasing the possibility of the customer receiving more interesting information, avoiding the customer's aversion to the material content sent by the channel, improving the customer's operation results, and helping to improve the customer's actual conversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0011] Figure 1 is a schematic diagram of an application environment of a customer channel matching method in one embodiment of the present invention;

[0012] Figure 2is a flow chart of a customer channel matching method in one embodiment of the present invention;

[0013] Figure 3 is a structural schematic diagram of a customer channel matching device in one embodiment of the present invention;

[0014] Figure 4 is a schematic diagram of a structure of a computer device in one embodiment of the present invention;

[0015] Figure 5 It is another structural schematic diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] The embodiment of the present application provides a customer channel matching method, which can be applied to a terminal, a server, or software running in a terminal or a server. In some embodiments, the terminal can be an electronic device such as a smart phone, a tablet computer, a laptop computer, or a desktop computer; the server can be configured as an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, and big data and artificial intelligence platforms; the software can be an application for customer channel matching, etc., but is not limited to the above forms.

[0018] For example, see Figure 1 Taking the customer channel matching method executed by an electronic device as an example, the electronic device can generate a comprehensive portrait feature of the customer to be matched by obtaining the customer portrait data and the customer channel behavior data of the customer to be matched, and input the comprehensive portrait feature of the customer to be matched into a pre-trained preset matching model to obtain a recommended marketing channel for the customer to be matched, so as to match the appropriate marketing channel for the customer to be matched, and complete the matching of the marketing channel for the customer to be matched. In this way, by selecting a suitable marketing channel through the pre-trained preset matching model for the unique customer data of different customers, or the unique customer group data for different customer groups, that is, the customer portrait data and the customer channel behavior data mentioned in the aforementioned embodiment, a more matching marketing channel can be provided to the customer, thereby improving the customer's operational effectiveness and helping to improve the actual conversion of the customer.

[0019] The present invention is described in detail below through specific embodiments. Figure 2 As shown, Figure 2 A flow chart of a customer channel matching method provided by an embodiment of the present invention includes the following steps:

[0020] Step S210, obtaining the customer portrait data and channel behavior data of the customer to be matched.

[0021] In one embodiment, the customer to be matched may be a specific customer or a typical representative of a customer group. The typical representative of the customer group may be a customer selected from multiple customers in the customer group as a typical representative.

[0022] It should be noted that the user data, channel data and other related data involved in the embodiments of the present application are collected with the authorization of users, enterprises and other related entities, or are actively submitted by users, enterprises and other related entities after reading the relevant instructions, or are data that users and other related entities need to upload when using the corresponding application. The collection of personal data and channel data in the embodiments of the present application is carried out under the premise of complying with relevant laws and regulations.

[0023] In this implementation, the customer portrait data to be matched and the sample customer portrait data mentioned in subsequent embodiments include but are not limited to customer classification, customer gender, etc. Customer classification can be the classification of customers by technical personnel within financial institutions such as banks.

[0024] In one embodiment, the customer channel behavior data to be matched and the sample customer channel behavior data mentioned in the subsequent embodiments include but are not limited to the following aspects:

[0025] 1) Behavior result data: that is, the customer’s historical behavior data, such as clicking on products, browsing product detail pages, purchasing products, etc.

[0026] 2) Behavioral intention data: This refers to the relevant data of customers before making a decision. The conversion rate will be higher if we use behavioral intention data to find people, such as searching for related products, paying attention to related information, and participating in related topics.

[0027] 3) Behavioral preference data: People with certain preferences or belonging to a certain type tend to be more inclined to buy a certain product. For example: AUM (Asset under Management) financial management preference, LUM (Liability under Management) loan preference, commodity coupon preference, etc., where LUM is mainly retail loans issued to customers, and AUM is the on- and off-balance sheet financial assets of retail customers including deposits, financial management, etc.

[0028] 4) Behavior pattern data: This is to summarize the behavioral performance with certain rules by analyzing the relationship between consumer behavior and time and space, as well as the spatial and temporal relationship between a series of behaviors, and predict related behaviors through these behavioral sequences. For example: pop-up behavior pattern, waterfall behavior pattern, 8 pm browsing behavior pattern, etc.

[0029] By analyzing the customer portrait data and channel behavior data of the customers to be matched, we can more clearly understand the characteristics of the customers to be matched, so as to provide more targeted channel recommendations for them in the future.

[0030] Step S220, generating comprehensive profile features of the customer to be matched based on the customer profile data to be matched and the customer channel behavior data to be matched.

[0031] Since the customer portrait data to be matched and the customer channel behavior data to be matched obtained in the above embodiment are scattered data, which is not conducive to the subsequent model input, it is necessary to process the above data. An exemplary method for generating comprehensive portrait features of customers to be matched, that is, generating comprehensive portrait features of customers to be matched according to the customer portrait data to be matched and the customer channel behavior data to be matched of customers to be matched, includes: generating customer portrait sub-features to be matched according to the customer portrait data to be matched; generating customer channel behavior sub-features to be matched according to the customer channel behavior data to be matched; generating comprehensive portrait features of customers to be matched based on customer portrait sub-features to be matched and customer channel behavior sub-features to be matched. By generating respective sub-features from the customer portrait data and channel behavior data, and then merging the two parts of sub-features to obtain customer comprehensive portrait features, the relevant data of customers to be matched can be vectorized so as to be input into the corresponding preset matching model for processing. It should be noted that the generation methods of comprehensive portrait features of customers to be matched and comprehensive portrait features of sample customers in this embodiment are similar. The specific generation method of comprehensive portrait features of customers to be matched can also refer to the following description of comprehensive portrait features of sample customers, which will not be repeated here.

[0032] Step S230, input the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain the recommended marketing channel for the customer to be matched, so as to serve as a suitable marketing channel for matching the customer to be matched.

[0033] Among them, the pre-trained preset matching model is obtained by training the improved boosting decision tree algorithm model by constructing a positive and negative sample sequence matrix through the historical conversion data of multiple sample customers in multiple marketing channels.

[0034] In one embodiment, before the comprehensive profile features of the customer to be matched are input into a pre-trained preset matching model, the method includes: obtaining sample customer portrait data of each sample customer, sample customer channel behavior data of each sample customer, sample channel portrait data of each sample marketing channel, and historical conversion results, wherein the historical conversion results characterize the conversion success or failure status of each sample customer in at least one sample marketing channel; generating sample customer comprehensive profile features of the target customer based on the sample customer portrait data of the target customer and the sample customer channel behavior data of the sample customer, wherein the target customer is a sample customer; generating sample channel portrait features based on the sample channel portrait data of each sample marketing channel; constructing a positive and negative sample sequence matrix based on the historical conversion results, all the sample channel portrait features, and all the sample customer comprehensive portrait features; and training the improved boosting decision tree algorithm model through the positive and negative sample sequence matrix to obtain a preset matching model.

[0035] The sample customer profile data and sample customer channel behavior data of the sample customers can be obtained through the customer data accumulation of the bank itself, or through other legal channels. The number of sample customers can be obtained by sampling data of customers with different customer profiles to ensure their richness. In this embodiment, when the sample customer profile data and sample customer channel behavior data of a certain customer exist in the data, the customer is used as a sample customer. In order to make the trained model more reliable, a sufficient number of sample customer data can be collected for model training. In this embodiment, the number of sample marketing channels is at least two, and the marketing channels that can be selected by the customers to be matched are all sample channels.

[0036] In one embodiment, generating a sample channel portrait feature of a sample marketing channel based on the sample channel portrait data of the sample marketing channel includes: generating a sample channel portrait subsequence of the target marketing channel based on at least one of the number of active customers, active moments, historical contact success rate, historical conversion success rate, number of responses at different moments, and number of responding customers at different moments of the target marketing channel, wherein the number of active customers, active moments, historical contact success rate, historical conversion success rate, number of responses at different moments, and number of responding customers at different moments are obtained based on the sample channel portrait data of the target marketing channel, and the target marketing channel is one of the sample marketing channels; constructing a sample channel portrait sequence based on the sample channel portrait subsequences generated by all the sample marketing channels; and determining the sample channel portrait sequence as the sample channel portrait feature. In other words, the sample channel portrait data of each marketing channel is serialized to obtain a sample channel portrait subsequence corresponding to the marketing channel, and the subsequences of all the sample marketing channels are integrated to obtain a sample channel portrait sequence, and the sequence is used as the sample channel portrait feature. It should be noted that for different sample channels, the types of sample channel portrait data may be the same or different. In this case, the sample channel portrait data of different sample channels may be cleaned and aligned by methods known to technical personnel in this field so that they can meet the needs of subsequent data processing.

[0037] In one embodiment, the sample marketing channels include but are not limited to at least one of APP (Application, marketing-related applications), WeChat, SMS, AI (Artificial Intelligence) outbound calls, etc.

[0038] Among them, the sample channel portrait data can be obtained by collecting the sample channel portrait data of each sample marketing channel based on a fixed time period, such as a week. The number of active customers can be based on certain statistical rules, such as the number of users who open the APP interface and stay for a preset stay time. The active time can be a certain time period, such as the answering rate of AI outbound calls is the highest from 8 to 9 pm. The historical contact success rate, that is, the proportion of customers who successfully receive marketing information to the total number of marketing customers in a fixed time period, such as the proportion of customers who answer AI outbound calls to the preset answering time to the total number of customers of AI outbound calls. The historical conversion success rate can be the proportion of the number of customers who successfully purchase or place orders for marketing target products to the total number of marketing customers (or the number of successfully contacted customers), which can be specifically set by technicians in this field as needed. The number of responses at different times and the number of responding customers at different times can be the number of responses (customer repeated counts) and the number of customers (the same customer is recorded as 1) in each sub-time period by dividing the fixed time period into different sub-time periods.

[0039] An example sample channel portrait sequence is S ij (S 1j ,S 2j …), where i is the channel feature identifier, the maximum value of i is the number of channel features (that is, the number of active customers, active moments, historical contact success rate, historical conversion success rate, number of responses at different times, and number of responding customers at different times that exist in the sample channel portrait data), j is the target channel identifier, and the maximum value of j is the number of sample channels.

[0040] In one embodiment, the sample customer comprehensive portrait features are generated based on the sample customer portrait data and the sample customer channel behavior data of the sample customer, including: generating sample customer portrait sub-features based on the sample customer portrait data; generating sample customer channel behavior sub-features based on the sample customer channel behavior data; generating sample customer comprehensive portrait features based on the sample customer portrait sub-features and the sample customer channel behavior sub-features. By generating respective sub-features from the customer's portrait data and channel behavior data, and then fusing the two sub-features to obtain the customer comprehensive portrait features, the customer's own objective situation and subjective preferences can be fully considered, and the subjective preferences (that is, customer channel behavior data) and traditional marketing recommendations based on objective age and gender are added, so that the real needs of customers can be understood more targetedly, and the acceptance of marketing can be further improved.

[0041] In this embodiment, sample customer channel behavior sub-features are generated based on sample customer channel behavior data, including: generating a clicked product sequence based on at least one of the sample customer's product clicking behavior data, product detail page browsing behavior data, and product purchasing behavior data; generating a behavior type preference sequence based on the sample customer's behavior preference data; generating a behavior category preference sequence based on at least one of the sample customer's related product search data, information attention data, and topic participation data; generating a position preference sequence based on at least one of the sample customer's customer behavior data, customer behavior time data, customer behavior space data, and customer behavior spatiotemporal relationship data; constructing sample customer weighted behavior sequence features based on the clicked product sequence, behavior type preference sequence, category preference sequence, and position preference sequence; and vectorizing the sample customer's weighted behavior sequence features to obtain sample customer channel behavior sub-features.

[0042] For example, a sequence of identifications of multiple products can be generated as a clicked product sequence according to the order of the products clicked. For another example, a clicked product sequence can be generated according to the product identifications of multiple products clicked by sample customers in sequence, whether the product details page of the product was browsed each time, the length of time spent on the details page, whether the browsed product was purchased, and other related data. i (i=0...i), where, I iThe i-th clicked product, or the i-th ranked product among multiple products clicked within a certain period of time, or the i-th ranked product is scored based on a comprehensive score based on the number of product views, the length of time spent on the details page, whether the viewed product was purchased, and other relevant data.

[0043] For example, by collecting statistics on the search results of sample customers for different types of financial products within a certain period of time, we can generate preference rankings for different types of financial products based on the products, consultations, topics, etc. that they searched for, and then obtain the customer preference rankings for different types of goods as the category preference sequence C. k (k=0…k), where C k is the type of financial product ranked kth in the preference ranking.

[0044] The behavior type preference sequence A can be determined based on the behavior preference data j (j=0…j), where A j The behavior preference ranked as the jth customer behavior type preference.

[0045] In this embodiment, a weighted behavior sequence feature of a sample customer is constructed based on a clicked product sequence, a behavior type preference sequence, a category preference sequence, and a position preference sequence; the weighted behavior sequence feature of the sample customer is vectorized to obtain a channel behavior sub-feature of the sample customer, including: weighted fusion of the sequences obtained in the above embodiments to construct a user weighted behavior sequence feature (that is, a sample customer channel behavior sub-feature) F2, and vectorization.

[0046] A positive sample sequence matrix is ​​constructed based on the relevant sample customer portrait data with successful historical conversion results, the sample customer channel behavior data, and the sample channel portrait data of successful sample marketing channels; a negative sample sequence matrix is ​​constructed based on the relevant sample customer portrait data with failed historical conversion results, the sample customer channel behavior data, and the sample channel portrait data of successful sample marketing channels; the negative sample sequence matrix and the positive sample sequence matrix are used as positive and negative sample sequence matrices to train the improved boosting decision tree algorithm model through the positive and negative sample sequence matrices to obtain a preset matching model.

[0047] Among them, the improved boosting decision tree algorithm model includes but is not limited to the XGBoost (eXtreme GradientBoosting) tree model, etc. The adaptability of different channels of customers can be predicted and scored (such as a percentage system, etc.). Multiple hyperparameters of the model, such as the learning rate, etc., can be set in advance.

[0048] In one embodiment, comprehensive profile features of customers to be matched are input into a pre-trained preset matching model to obtain recommended marketing channels for customers to be matched, including: inputting comprehensive profile features of customers to be matched into a pre-trained preset matching model to obtain a matching score for each marketing channel by the preset matching model; ranking multiple marketing channels from low to high scores based on the matching scores; and determining the marketing channel with the highest ranking as the recommended marketing channel.

[0049] That is, through the pre-trained preset matching model, the matching situation between the customers to be matched and different marketing channels can be scored, and then the recommended marketing channels can be determined according to the scores of different marketing channels, which can make the marketing to customers more targeted and further increase the possibility of customer satisfaction with marketing.

[0050] In one embodiment, comprehensive profile features of customers to be matched are input into a pre-trained preset matching model to obtain recommended marketing channels for customers to be matched, including: inputting comprehensive profile features of customers to be matched into a pre-trained preset matching model to obtain a matching score for each marketing channel by the preset matching model; if there are at least two preferred channels, the preferred channel is a marketing channel whose matching score is greater than a preset score threshold, and the channel marketing cost of each preferred channel is obtained; and the preferred channel with the smallest channel marketing cost is determined as the recommended marketing channel.

[0051] Sometimes, a certain customer can be adapted to at least two marketing channels. At this time, the appropriate marketing channel can be selected as the recommended marketing channel for the customer based on the current cost of different marketing channels, so as to market to the customer through the recommended marketing channel. For example, the current marketing cost of a single customer for AI outbound calling is 50 yuan, and AI outbound calling does not have a certain package volume available. At this time, the channel marketing fee for AI outbound calling is 50 yuan, and the current package usage of SMS marketing has not exceeded the package limit. It can be considered that its additional marketing cost is 0 yuan. If the customer can accept both AI outbound calling and SMS marketing, and the matching scores are high, the SMS marketing channel can be selected based on the principle of minimum call charges.

[0052] The channel marketing expenses may be determined by a unified standard pre-set by those skilled in the art, or may be determined in real time based on the current number of marketing customers of different marketing channels, the remaining amount of marketing expenses, etc., or by other methods known to those skilled in the art.

[0053] Below, through a specific embodiment, the customer channel matching method provided in the above embodiment is exemplarily described. First, for the customer to be matched, the age, gender, customer group classification, and previously collected channel behavior data such as click product status, browse product status, and search product status are obtained to generate a comprehensive portrait feature of the customer to be matched, and the comprehensive portrait feature of the customer to be matched is input into the preset matching model, so that the preset matching model can score multiple marketing channels based on the comprehensive portrait feature of the customer to be matched, and select a marketing channel as a recommended marketing channel based on the score, so as to realize the construction of customer portraits and the mining of user behavior sequence data of different channels, correctly match the preference of customers and channels, find the most suitable contact conversion channel for customers, and find the most suitable audience customer groups for different channels, so as to improve customer operation results and be more conducive to improving the actual conversion of customers.

[0054] It can be seen that in the above scheme, by obtaining the customer portrait data and channel behavior data of the customer to be matched, a comprehensive portrait feature of the customer to be matched is generated, and the comprehensive portrait feature of the customer to be matched is input into the pre-trained preset matching model to obtain the recommended marketing channel for the customer to be matched, so as to match the appropriate marketing channel for the customer to be matched, and complete the matching of the marketing channel for the customer to be matched. In this way, by using the unique customer data of different customers, or the unique customer group data of different customer groups, that is, the customer portrait data and channel behavior data to be matched mentioned in the aforementioned embodiment, a pre-trained preset matching model is used to select a suitable marketing channel, which can provide customers with more matching marketing channels, improve customer operation results, and help to improve the actual conversion of customers.

[0055] In traditional customer marketing plans, there is no in-depth understanding and exploration of the relationship between customers and channels, which leads to no consideration of customer feelings. Customers are blindly selected according to business rules and the same popular channel content is used to reach customers. As a result, customers cannot receive information they are interested in, are disgusted with the content sent by the channel, and even report it, resulting in a very poor user experience. It also causes a waste of operating resources and reduces customer stickiness and loyalty.

[0056] The inventors found that due to different business scenarios, the conversion rate, reach rate and response rate of different customer groups in different channels are different. In the retail customer marketing model of banks, it is an important way to improve customer experience and conversion to operate customers in a personalized manner based on their historical behavior data. With the construction of customer portraits by Internet technology and the mining of user behavior sequence data in different channels, the preferences of customers and channels are correctly matched, the most suitable reach and conversion channels for customers are found, and the most suitable audience groups are found for different channels, so as to improve customer operation results and improve the actual conversion of customers.

[0057] The unified operation and adaptive channel matching of customers in the basic retail field of the banking industry are important optimization and innovation difficulties. The customer channel matching method provided in the embodiment of the present application establishes an optimal solution model with the goal of successful conversion of customer channels on the basis of channel portraits, customer portraits and customer behavior sequence data of different channels, realizes adaptive matching of different customers and channels, reduces customer complaint rates, and greatly improves customer experience. It can also fully maximize the utilization of channel resources, reduce the waste of operating costs, and achieve a win-win situation for customers and businesses.

[0058] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0059] In one embodiment, a customer channel matching device is provided, which corresponds one-to-one to the customer channel matching method in the above embodiment. Figure 3 As shown, the customer channel matching device includes: an acquisition module 301, a generation module 302 and a matching module 303. The functional modules are described in detail as follows: an acquisition module, used to acquire the customer portrait data to be matched and the customer channel behavior data to be matched of the customer to be matched; a generation module, used to generate the comprehensive portrait features of the customer to be matched according to the customer portrait data to be matched and the customer channel behavior data to be matched; a matching module, used to input the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain the recommended marketing channel for the customer to be matched, so as to match the customer to be matched with a suitable marketing channel, wherein the pre-trained preset matching model is obtained by constructing a positive and negative sample sequence matrix through the historical conversion data of multiple sample customers in multiple marketing channels to train the improved boosting decision tree algorithm model.

[0060] In one embodiment, the device also includes a pre-training module, which is used to obtain sample customer portrait data of each sample customer, sample customer channel behavior data of each sample customer, sample channel portrait data of each sample marketing channel, and historical conversion results, wherein the historical conversion results characterize the conversion success or failure status of each sample customer in at least one sample marketing channel; generate sample customer comprehensive portrait features of the target customer based on the sample customer portrait data of the target customer and the sample customer channel behavior data of the sample customer, wherein the target customer is a sample customer; generate sample channel portrait features based on the sample channel portrait data of each sample marketing channel; construct a positive and negative sample sequence matrix based on the historical conversion results, all sample channel portrait features and all sample customer comprehensive portrait features; and train the improved boosting decision tree algorithm model through the positive and negative sample sequence matrix to obtain a preset matching model.

[0061] In one embodiment, the pre-training module is also used to: generate a sample channel portrait subsequence of the target marketing channel according to at least one of the number of active customers, active moments, historical contact success rate, historical conversion success rate, number of responses at different moments, and number of responding customers at different moments of the target marketing channel, the number of active customers, active moments, historical contact success rate, historical conversion success rate, number of responses at different moments, and number of responding customers at different moments are obtained based on sample channel portrait data of the target marketing channel, and the target marketing channel is a sample marketing channel; construct a sample channel portrait sequence based on the sample channel portrait subsequences generated by all the sample marketing channels; and determine the sample channel portrait sequence as a sample channel portrait feature.

[0062] In one embodiment, the pre-training module is also used to: generate sample customer portrait sub-features based on sample customer portrait data; generate sample customer channel behavior sub-features based on sample customer channel behavior data; and generate sample customer comprehensive portrait features based on sample customer portrait sub-features and sample customer channel behavior sub-features.

[0063] In one embodiment, the pre-training module is also used to: generate a clicked product sequence based on at least one of the sample customer's product clicking behavior data, product detail page browsing behavior data, and product purchasing behavior data; generate a behavior type preference sequence based on the sample customer's behavior preference data; generate a behavior category preference sequence based on at least one of the sample customer's related product search data, information attention data, and topic participation data; generate a location preference sequence based on at least one of the sample customer's customer behavior data, customer behavior time data, customer behavior space data, and customer behavior spatiotemporal relationship data; construct a weighted behavior sequence feature of the sample customer based on the clicked product sequence, behavior type preference sequence, category preference sequence, and location preference sequence; and vectorize the sample customer's weighted behavior sequence feature to obtain the sample customer's channel behavior sub-feature.

[0064] In one embodiment, the matching module is also used to: input the comprehensive profile features of the customer to be matched into a pre-trained preset matching model to obtain the matching score of the preset matching model for each marketing channel; sort multiple marketing channels from low to high scores based on the matching score; and determine the marketing channel with the highest ranking as the recommended marketing channel.

[0065] In one embodiment, the matching module is also used to: input the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain the matching score of the preset matching model for each marketing channel; if there are at least two preferred channels, the preferred channel is a marketing channel whose matching score is greater than a preset score threshold, and the channel marketing cost of each preferred channel is obtained; and the preferred channel with the smallest channel marketing cost is determined as the recommended marketing channel.

[0066] An embodiment of the present invention provides a customer channel matching device, which generates a comprehensive portrait feature of the customer to be matched by acquiring the customer portrait data and the customer channel behavior data of the customer to be matched, and inputs the comprehensive portrait feature of the customer to be matched into a pre-trained preset matching model to obtain a recommended marketing channel for the customer to be matched, so as to match the appropriate marketing channel for the customer to be matched, thereby completing the matching of the marketing channel for the customer to be matched. In this way, by using the unique customer data of different customers, or the unique customer group data of different customer groups, that is, the customer portrait data and the customer channel behavior data mentioned in the aforementioned embodiment, a pre-trained preset matching model is used to select a suitable marketing channel, which can provide the customer with a more matching marketing channel, improve the customer operation results, and help to improve the actual conversion of customers.

[0067] For the specific definition of the customer channel matching device, please refer to the definition of the customer channel matching method above, which will not be repeated here. Each module in the above customer channel matching device can be implemented in whole or in part by software, hardware and a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0068] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, a function or step on the service side of a customer channel matching method is implemented.

[0069] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server through a network connection. When the computer program is executed by the processor, it implements the functions or steps on the client side of a customer channel matching method.

[0070] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: obtaining customer portrait data and customer channel behavior data of the customer to be matched; generating comprehensive portrait features of the customer to be matched based on the customer portrait data and customer channel behavior data; inputting the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain recommended marketing channels for the customer to be matched, so as to serve as suitable marketing channels for matching the customer to be matched, wherein the pre-trained preset matching model is obtained by constructing a positive and negative sample sequence matrix based on historical conversion data of multiple sample customers in multiple marketing channels to train an improved boosting decision tree algorithm model.

[0071] The computer device provided in the above embodiment generates comprehensive portrait features of the customer to be matched by acquiring the customer portrait data and channel behavior data of the customer to be matched, and inputs the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain recommended marketing channels for the customer to be matched, so as to match the customer with a suitable marketing channel. By selecting a suitable marketing channel based on the unique customer data of different customers through a pre-trained preset matching model, the customer can be provided with a more matching marketing channel, thereby improving the customer's operational effectiveness and facilitating the improvement of the customer's actual conversion.

[0072] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: obtaining customer portrait data and customer channel behavior data of the customer to be matched; generating comprehensive portrait features of the customer to be matched based on the customer portrait data and customer channel behavior data; inputting the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain recommended marketing channels for the customer to be matched, so as to serve as suitable marketing channels for matching the customer to be matched, wherein the pre-trained preset matching model is obtained by constructing a positive and negative sample sequence matrix through historical conversion data of multiple sample customers in multiple marketing channels to train an improved boosting decision tree algorithm model.

[0073] The computer-readable storage medium provided in the above embodiment generates comprehensive portrait features of the customer to be matched by acquiring the customer portrait data and the customer channel behavior data of the customer to be matched, and inputs the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain recommended marketing channels for the customer to be matched, so as to match the customer with a suitable marketing channel. By selecting a suitable marketing channel based on the unique customer data of different customers through a pre-trained preset matching model, the customer can be provided with a more matching marketing channel, thereby improving the customer's operational effectiveness and facilitating the improvement of the customer's actual conversion.

[0074] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0075] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0076] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0077] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A customer channel matching method, characterized in that: The method comprises: Obtain the customer profile data and channel behavior data of the customer to be matched; Generate comprehensive profile features of the customer to be matched based on the customer profile data to be matched and the customer channel behavior data to be matched; The comprehensive portrait features of the customer to be matched are input into a pre-trained preset matching model to obtain the recommended marketing channels for the customer to be matched, so as to serve as the appropriate marketing channels for matching the customer to be matched, wherein the pre-trained preset matching model is obtained by training the improved boosting decision tree algorithm model by constructing a positive and negative sample sequence matrix through historical conversion data of multiple sample customers in multiple marketing channels.

2. The customer channel matching method according to claim 1, characterized in that: Before inputting the comprehensive profile features of the to-be-matched customer into a pre-trained preset matching model, the method includes: Obtaining sample customer portrait data of each sample customer, sample customer channel behavior data of each sample customer, sample channel portrait data of each sample marketing channel, and historical conversion results, wherein the historical conversion results represent the conversion success or failure status of each sample customer in at least one sample marketing channel; Generate a sample customer comprehensive portrait feature of the target customer based on the sample customer portrait data of the target customer and the sample customer channel behavior data of the sample customer, where the target customer is a sample customer; Generate sample channel profile features based on sample channel profile data of each sample marketing channel; Construct a positive and negative sample sequence matrix based on historical conversion results, all sample channel profile features, and all sample customer comprehensive profile features; The improved boosting decision tree algorithm model is trained by using the positive and negative sample sequence matrix to obtain a preset matching model.

3. The customer channel matching method according to claim 2, characterized in that: The sample channel portrait features of the sample marketing channel generated according to the sample channel portrait data of the sample marketing channel include: generating a sample channel portrait subsequence of the target marketing channel according to at least one of the number of active customers, active moments, historical contact success rate, historical conversion success rate, number of responses at different moments, and number of responding customers at different moments of the target marketing channel, wherein the number of active customers, active moments, historical contact success rate, historical conversion success rate, number of responses at different moments, and number of responding customers at different moments are obtained based on the sample channel portrait data of the target marketing channel, and the target marketing channel is a sample marketing channel; Constructing a sample channel portrait sequence based on the sample channel portrait subsequences generated by all sample marketing channels; The sample channel portrait sequence is determined as the sample channel portrait feature.

4. The customer channel matching method according to claim 2, characterized in that: Generate sample customer comprehensive profile features based on sample customer profile data and sample customer channel behavior data of the sample customers, including: Generate sample customer profile sub-features based on sample customer profile data; Generate sample customer channel behavior sub-features based on sample customer channel behavior data; The sample customer comprehensive portrait feature is generated based on the sample customer portrait sub-feature and the sample customer channel behavior sub-feature.

5. The customer channel matching method according to claim 2, characterized in that: Generate sample customer channel behavior sub-features based on sample customer channel behavior data, including: Generate a clicked product sequence based on at least one of the sample customers' clicked product behavior data, product detail page browsing behavior data, and product purchase behavior data; Generate a behavior type preference sequence based on the behavior preference data of sample customers; Generate a behavior category preference sequence based on at least one of the sample customers' related product search data, information attention data, and topic participation data; Generate a location preference sequence based on at least one of customer behavior data, customer behavior time data, customer behavior space data, and customer behavior spatiotemporal relationship data of sample customers; Constructing a weighted behavior sequence feature of sample customers based on the clicked product sequence, behavior type preference sequence, category preference sequence, and location preference sequence; Vectorization is performed on the weighted behavior sequence features of the sample customers to obtain channel behavior sub-features of the sample customers.

6. The customer channel matching method according to any one of claims 1 to 5, characterized in that: Input the comprehensive profile features of the customer to be matched into a pre-trained preset matching model to obtain the recommended marketing channels for the customer to be matched, including: Inputting the comprehensive profile features of the customer to be matched into a pre-trained preset matching model to obtain a matching score of each marketing channel by the preset matching model; sorting the plurality of marketing channels from low score to high score based on the matching score; The marketing channel with the highest ranking is determined as the recommended marketing channel.

7. The customer channel matching method according to any one of claims 1 to 5, characterized in that: Input the comprehensive profile features of the customer to be matched into a pre-trained preset matching model to obtain the recommended marketing channels for the customer to be matched, including: Inputting the comprehensive profile features of the customer to be matched into a pre-trained preset matching model to obtain a matching score of each marketing channel by the preset matching model; If there are at least two preferred channels, the preferred channels are marketing channels with matching scores greater than a preset score threshold, and the channel marketing costs of each preferred channel are obtained; The preferred channel with the lowest channel marketing cost is determined as the recommended marketing channel.

8. A customer channel matching device, characterized in that: The device comprises: An acquisition module is used to acquire the customer portrait data and channel behavior data of the customer to be matched; A generation module, used to generate comprehensive profile features of the customer to be matched based on the customer profile data to be matched and the customer channel behavior data to be matched; A matching module is used to input the comprehensive portrait features of the customer to be matched into a pre-trained preset matching model to obtain the recommended marketing channel for the customer to be matched, so as to serve as a suitable marketing channel for matching the customer to be matched, wherein the pre-trained preset matching model is obtained by training an improved boosting decision tree algorithm model by constructing a positive and negative sample sequence matrix through historical conversion data of multiple sample customers in multiple marketing channels.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the customer channel matching method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the customer channel matching method according to any one of claims 1 to 7 are implemented.

Images