Product push method and device

Abstract

The present invention provides a product push method and device, which belongs to the field of artificial intelligence technology and can be applied to the financial field or other fields. The product push method includes: obtaining customer preferred product information, generating customer ideal product data according to the customer preferred product information; generating product similarity of each financial product according to the customer ideal product data and each financial product data; inputting each product similarity and the marketing forecast data of each marketing personnel into a transaction scale model created based on product historical similarity and historical marketing data to obtain each transaction scale data; determining recommended products and push personnel according to each transaction scale data, pushing recommended products to customers and cutting the signal source of the push personnel into the customer's account. The present invention can meet the real demands of customers and improve the accuracy of product push.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a product push method and device. Background Art

[0002] In the field of bank agency sales, traditional banks manually compile sales plans according to the strength of sales promotion, first recommend the main products, and then push them to potential customers indiscriminately according to the sales plan. Currently, all banks are exploring the use of historical data to build customer portraits, establish matching relationships between customer portraits and products, and use big data analysis to form push results that are tailored to each customer.

[0003] The recommended products of traditional banks are too fixed and are completely based on the bank's perspective, ignoring the differences of customers. The push methods of various banks based on customer portraits also have the following problems:

[0004] 1. Data has problems such as low availability and untimely updates, which affect the construction of customer portraits and, in turn, the accuracy of product push.

[0005] 2. The push results lack feedback and correction mechanisms. Only successful push results can enter the database and become historical data. The construction of the push method is too one-sided.

[0006] 3. The influence of push personnel on customer decision-making is ignored, and the push results cannot accurately reflect the customer’s true intentions. Summary of the invention

[0007] The main purpose of the embodiments of the present invention is to provide a product push method and device to meet the real demands of customers and improve the accuracy of product push.

[0008] In order to achieve the above object, an embodiment of the present invention provides a product push method, including:

[0009] Obtain information about products preferred by customers, and generate data about customers' ideal products based on the information;

[0010] Generate product similarity of each financial product based on the customer's ideal product data and each financial product data;

[0011] Input the similarity of each product and the marketing forecast data of each marketing personnel into the transaction scale model created based on the historical similarity of the product and the historical marketing data to obtain the transaction scale data;

[0012] Determine recommended products and push personnel based on transaction size data, push recommended products to customers and connect the push personnel's signal source to the customer's account.

[0013] The embodiment of the present invention further provides a product pushing device including:

[0014] An ideal product data generation module is used to obtain customer preferred product information and generate customer ideal product data based on the customer preferred product information;

[0015] The product similarity module is used to generate the product similarity of each financial product based on the customer's ideal product data and the data of each financial product;

[0016] The transaction scale data module is used to input the similarity of each product and the marketing forecast data of each marketing personnel into the transaction scale model created based on the historical similarity of the product and the historical marketing data to obtain the transaction scale data;

[0017] The push module is used to determine the recommended products and push personnel based on the transaction scale data, push the recommended products to the customers and cut the push personnel's signal source into the customer's account.

[0018] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements the steps of the product push method when executing the computer program.

[0019] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the product push method are implemented.

[0020] The product push method and device of the embodiment of the present invention first generates the similarity of each product based on the data of each financial product and the customer's ideal product data generated by the customer's product preference information, and then generates each transaction scale data based on the similarity of each product and the marketing forecast data of each marketer to determine the recommended products and push personnel and push them to the customer. This can meet the real demands of the customer and improve the accuracy of product push. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 is a flow chart of a product push method in one embodiment of the present invention;

[0023] Figure 2 is a flow chart of a product push method in another embodiment of the present invention;

[0024] Figure 3 is a flow chart of S102 in one embodiment of the present invention;

[0025] Figure 4 is a flow chart of determining customer expression percentage data in an embodiment of the present invention;

[0026] Figure 5 is a flow chart of determining the customer's face facing data in an embodiment of the present invention;

[0027] Figure 6 is a flow chart of creating a transaction scale model in an embodiment of the present invention;

[0028] Figure 7 It is a structural block diagram of a product pushing device in an embodiment of the present invention;

[0029] Figure 8 It is a structural block diagram of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] 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 only 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.

[0031] Those skilled in the art will appreciate that the embodiments of the present invention may be implemented as a system, device, apparatus, method or computer program product. Therefore, the present disclosure may be implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0032] In view of the fact that the product push results of the prior art cannot accurately reflect the true intentions of customers, which affects the accuracy of product push, the embodiment of the present invention provides a product push method, which uses natural language processing and facial expression recognition technology to analyze the customer's expression language and body language during the marketing push process, establishes a connection with historical data to form a model, and updates the model iteratively through machine learning technology, and finally compares the output results to select the optimal input value to build a recommendation system that can close the marketing process, meet the real demands of customers, and improve the accuracy of product push. The present invention is described in detail below in conjunction with the accompanying drawings.

[0033] Figure 1 It is a flow chart of a product push method in one embodiment of the present invention. Figure 2 FIG. 1 is a flow chart of a product push method in another embodiment of the present invention. Figure 1 and Figure 2 As shown, the product push methods include:

[0034] S101: Acquire customer preferred product information, and generate customer ideal product data based on the customer preferred product information.

[0035] Before executing S101, the method further includes: obtaining the voice information of the customer during the historical product push period and converting the voice information into text. Then, the customer's focus on the product (customer preferred product information) is extracted from the text through natural language processing keyword extraction technology to construct the customer's ideal product.

[0036] Among them, the customer's ideal product data includes risk data x￠ (dimensionless grade data, with a value of 1-5), return data y￠ (annualized rate of return, in %) and term data z￠ (in days). A three-dimensional coordinate system is constructed from the three dimensions of risk data, return data and term data, and the customer's ideal product a￠ (x￠, y￠, z￠) and the i-th financial product a i (x i ,y i ,z i ) are substituted into the coordinate system to generate space vectors oa￠ and oa i .

[0037] S102: Generate product similarity of each financial product based on the customer's ideal product data and each financial product data.

[0038] Figure 3 1 is a flow chart of S102 in one embodiment of the present invention. Figure 3 As shown, S102 includes:

[0039] S201: Determine the Euclidean distance of each product based on the customer's ideal product data and each financial product.

[0040] In specific implementation, the Euclidean distance of each product can be determined by the following formula:

[0041]

[0042] Among them, r i is the Euclidean distance of the ith product, x i is the risk data of the ith financial product, y i is the income data of the ith financial product, z i is the term data of the ith financial product, x￠ is the risk data of the customer’s ideal product, y￠ is the return data of the customer’s ideal product, and z￠ is the term data of the customer’s ideal product.

[0043] S202: Determine the product similarity of each financial product based on the Euclidean distance of each product.

[0044] In specific implementation, the product similarity of each financial product can be determined by the following formula:

[0045]

[0046] Among them, A i is the product similarity of the ith financial product. The value range of product similarity is [0,1]. The smaller the Euclidean distance of the product, the greater the product similarity.

[0047] S103: Inputting the similarity of each product and the marketing forecast data of each marketing personnel into a transaction scale model created based on the historical similarity of the products and the historical marketing data to obtain each transaction scale data.

[0048] The transaction scale data is linearly related to product similarity and marketing prediction data of marketers. The transaction scale data corresponding to the i-th financial product recommended by the k-th marketer can be obtained according to the following formula:

[0049] X ki =a0+a1A i +a2B k ;

[0050] Among them, X ki The transaction scale data corresponding to the recommendation of the i-th financial product to the k-th marketing personnel, a0 is the first scale model parameter, a1 is the second scale model parameter, a2 is the third scale model parameter, A i is the product similarity of the ith financial product, B k is the marketing prediction data corresponding to the k-th marketing personnel.

[0051] S104: Determine the recommended products and push personnel according to the transaction scale data, push the recommended products to the customer and insert the push personnel's signal source into the customer's account.

[0052] In one embodiment, determining the recommended products and push personnel according to each transaction scale data includes:

[0053] Determine the financial product corresponding to the maximum value of each transaction scale data as the recommended product, and determine the maximum value of each transaction scale data X ki The marketing personnel corresponding to the maximum value of is the push personnel.

[0054] For example, transaction size data X ki The maximum value of X 23 , then push the third financial product to the customer. Cut the signal source of the second marketing staff into the customer's account, so that the second marketing staff can market the third financial product to the customer.

[0055] Figure 1 The execution subject of the product push method shown can be a computer. Figure 1As can be seen from the process shown, the product push method of the embodiment of the present invention first generates the similarity of each product based on the data of each financial product and the customer's ideal product data generated by the customer's product preference information, and then generates each transaction scale data based on the similarity of each product and the marketing forecast data of each marketer to determine the recommended products and push personnel and push them to the customer, which can meet the real demands of the customer and improve the accuracy of product push.

[0056] In one embodiment, the product push method further includes:

[0057] 1. Obtain the transaction behavior data and customer facial data corresponding to each marketing personnel.

[0058] Among them, customer face data includes customer expression ratio data n1 and customer face facing data n2. Transaction behavior data includes transaction completion data n3, order withdrawal data after completion n4, customer response data n5 during historical product push, customer repurchase data n6, transaction amount as a percentage of customer account active cash assets data n7 and product sharing data n8. Response data n5 can be the ratio of the number of customer responses to the number of explanations by the push personnel during the historical product push.

[0059] 2. Input the transaction behavior data and customer face data corresponding to each marketing personnel into the marketing prediction model created based on the transaction behavior training data and customer face training data to obtain the marketing prediction data of each marketing personnel.

[0060] In specific implementation, the marketing prediction model can adopt the XGBOOST model. The initial XGBOOST model is trained according to the transaction behavior training data (historical values ​​of transaction behavior data) and customer face training data (historical values ​​of customer face data). The leaf node parameters l of the CART tree in the XGBOOST model can be obtained. j , according to the parameters l of each leaf node j Create a marketing forecasting model. j = 1, 2, 3, 4, 5, 6, 7, 8.

[0061] The indicators corresponding to the marketing personnel (including customer expression ratio data n1, customer face data n2, transaction completion data n3, order cancellation data after completion n4, customer response data during historical product push n5, customer repurchase data n6, transaction amount as a percentage of customer account active cash assets data n7 and product sharing data n8) are regarded as leaf nodes of the CART tree, and the data f corresponding to the jth leaf node is calculated. j (n) = l j n j , the corresponding data of each leaf node can be accumulated to obtain the marketing forecast data corresponding to the marketing personnel Then the marketing forecast data corresponding to the kth marketing personnel is B k .

[0062] Figure 4 1 is a flow chart of determining customer expression percentage data in an embodiment of the present invention. Figure 4 As shown, the data for determining the proportion of customer expressions includes:

[0063] S301: Acquire the customer's historical video, and acquire the customer's facial expression image based on the customer's historical video.

[0064] During specific implementation, the customer's historical video may be screenshoted to obtain the customer's facial expression image.

[0065] S302: Using a convolutional neural network to extract features from customer expression images to obtain the probability of each customer expression.

[0066] In specific implementation, convolutional neural networks and facial expression recognition technology can be used to extract features from customer expression images. The convolutional neural network model outputs the probability distribution of preset classification expressions, such as happy expressions, neutral expressions, disgusted expressions, and puzzled expressions.

[0067] S303: Determine customer expression proportion data according to the expression probability of each customer.

[0068] During specific implementation, the probability of the customer's happy expression appearing can be determined as the customer's expression proportion data n1.

[0069] Figure 5 1 is a flow chart of determining the customer's face facing data in an embodiment of the present invention. Figure 5 As shown, the data for determining the customer's face facing includes:

[0070] S401: Acquire a customer's facial image based on the customer's historical video.

[0071] During specific implementation, the customer's historical video can be screenshoted to obtain the customer's facial image.

[0072] S402: Inputting the customer's face image into a face angle recognition model created based on the face training image to obtain the customer's face angle data.

[0073] In specific implementation, the face images facing the screen at different angles can be used as face training images to train the initial YOLO-V3 (the third version of the YOLO (You Only Look Once) series target detection algorithm) model to create a face angle recognition model. Face training images include training images facing the screen, training images offset by 60 degrees from the screen, and training images offset by 30 degrees from the screen. After the training is completed, each customer's face image is input into the face angle recognition model to obtain the customer's face angle data corresponding to each customer's face image; the customer's face angle data includes data facing the screen, data offset by 60 degrees from the screen, and data offset by 30 degrees from the screen.

[0074] S403: Determine the customer's face facing data according to the customer's face angle data.

[0075] During specific implementation, the proportion of the data facing the screen to the customer's face angle data, i.e., the customer's face facing data n2, can be determined.

[0076] Figure 6 1 is a flow chart of creating a transaction scale model in an embodiment of the present invention. Figure 6 As shown, creating a transaction size model based on product historical similarity and historical marketing data includes:

[0077] Perform the following iterative processing:

[0078] S501: Determine scale prediction data based on product historical similarity, historical marketing data, and scale model parameters.

[0079] In specific implementation, the product historical similarity A￠ and historical marketing data B￠ can be input into the initial transaction scale model to obtain the scale prediction data X￠, as follows:

[0080] X￠=a0+a1A￠+a2B￠;

[0081] Among them, a0 is the first scale model parameter, a1 is the second scale model parameter, a2 is the third scale model parameter, and the scale model parameter a={a0, a1, a2}.

[0082] S502: Determine a scale loss function according to the scale prediction data and the corresponding scale actual data.

[0083] S503: Determine whether the scale loss function is less than a preset scale loss function threshold.

[0084] S504: When the scale loss function is less than a preset scale loss function threshold, a transaction scale model is created according to the scale model parameters.

[0085] S505: When the scale loss function is greater than or equal to a preset scale loss function threshold, the scale model parameters are updated according to the scale loss function, and the iterative process is continued.

[0086] Among them, the scale loss function can be processed according to the gradient descent algorithm to update the scale model parameters.

[0087] The specific process of the embodiment of the present invention is as follows:

[0088] 1. Determine scale forecast data based on product historical similarity, historical marketing data and scale model parameters.

[0089] 2. Determine the scale loss function based on the scale prediction data and the corresponding scale actual data.

[0090] 3. When the scale loss function is less than the preset scale loss function threshold, create a transaction scale model according to the scale model parameters; otherwise, update the scale model parameters according to the scale loss function and return to step 1.

[0091] 4. Obtain information about products preferred by customers, and generate data about customers’ ideal products based on the information.

[0092] 5. Determine the Euclidean distance of each product based on the customer’s ideal product data and each financial product.

[0093] 6. Determine the product similarity of each financial product based on the Euclidean distance of each product.

[0094] 7. Obtain the customer's historical videos, and obtain the customer's expression images and customer's face images based on the customer's historical videos.

[0095] 8. Use convolutional neural networks to extract features from customer expression images to obtain the probability of each customer’s expression, and determine the customer’s expression percentage data based on the probability of each customer’s expression.

[0096] 9. Input the customer's facial image into the facial angle recognition model created based on the facial training image to obtain the customer's facial angle data.

[0097] 10. Determine the customer’s face facing data based on the customer’s face angle data.

[0098] 11. Obtain the transaction behavior data corresponding to each marketing personnel, and input the transaction behavior data corresponding to each marketing personnel and the customer face data including customer expression ratio data and customer face facing data into the marketing prediction model created based on the transaction behavior training data and the customer face training data to obtain the marketing prediction data of each marketing personnel.

[0099] 12. Input the transaction size model based on the similarity of each product and the marketing forecast data of each marketer to obtain the transaction size data.

[0100] 13. Determine the recommended products and push personnel based on the transaction scale data, push the recommended products to the customers and cut the push personnel's signal source into the customer's account.

[0101] In summary, the product push method provided in the embodiment of the present invention uses artificial intelligence and machine learning to establish a relationship between the information conveyed by the customer's verbal expression and body language and the recommended products and push personnel, and provides a recommendation system that is more in line with customer demands and can be self-iteratively optimized. On the one hand, it pays attention to the fact that changes in various external environmental factors may affect the accuracy of the push method, and the deviation can be effectively corrected through the continuous iteration and optimization of the model parameter a; on the other hand, it deeply mines and analyzes the marketing process to obtain quantifiable real customer demands, establishes a model to correct the push results, and achieves a win-win situation for customers and banks.

[0102] Based on the same inventive concept, an embodiment of the present invention further provides a product pushing device. Since the principle of solving the problem by the device is similar to that of the product pushing method, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0103] Figure 7 is a structural block diagram of a product pushing device in an embodiment of the present invention. Figure 7 As shown, the product pushing device includes:

[0104] An ideal product data generation module is used to obtain customer preferred product information and generate customer ideal product data based on the customer preferred product information;

[0105] The product similarity module is used to generate the product similarity of each financial product based on the customer's ideal product data and the data of each financial product;

[0106] The transaction scale data module is used to input the similarity of each product and the marketing forecast data of each marketing personnel into the transaction scale model created based on the historical similarity of the product and the historical marketing data to obtain the transaction scale data;

[0107] The push module is used to determine the recommended products and push personnel based on the transaction scale data, push the recommended products to the customers and cut the push personnel's signal source into the customer's account.

[0108] In summary, the product push device in the embodiment of the present invention first generates the similarities of each product based on the data of each financial product and the customer's ideal product data generated by the customer's product preference information, and then generates each transaction scale data based on the similarities of each product and the marketing forecast data of each marketer to determine the recommended products and push personnel and push them to the customer, which can meet the real demands of the customer and improve the accuracy of product push.

[0109] The embodiment of the present invention also provides a specific implementation of a computer device that can implement all the steps in the product push method in the above embodiment. Figure 8 is a block diagram of a computer device in an embodiment of the present invention, see Figure 8 , the computer equipment specifically includes the following contents:

[0110] Processor (processor) 801 and memory (memory) 802.

[0111] The processor 801 is used to call the computer program in the memory 802. When the processor executes the computer program, all steps in the product push method in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0112] Obtain information about products preferred by customers, and generate data about customers' ideal products based on the information;

[0113] Generate product similarity of each financial product based on the customer's ideal product data and each financial product data;

[0114] Input the similarity of each product and the marketing forecast data of each marketing personnel into the transaction scale model created based on the historical similarity of the product and the historical marketing data to obtain the transaction scale data;

[0115] Determine recommended products and push personnel based on transaction size data, push recommended products to customers and connect the push personnel's signal source to the customer's account.

[0116] In summary, the computer device of the embodiment of the present invention first generates the similarities of each product based on the data of each financial product and the customer's ideal product data generated by the customer's preferred product information, and then generates each transaction scale data based on the similarities of each product and the marketing forecast data of each marketer to determine the recommended products and push personnel and push them to the customer, which can meet the real demands of the customer and improve the accuracy of product push.

[0117] The embodiment of the present invention also provides a computer-readable storage medium capable of implementing all the steps of the product push method in the above embodiment. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, all the steps of the product push method in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0118] Obtain information about products preferred by customers, and generate data about customers' ideal products based on the information;

[0119] Generate product similarity of each financial product based on the customer's ideal product data and each financial product data;

[0120] Input the similarity of each product and the marketing forecast data of each marketing personnel into the transaction scale model created based on the historical similarity of the product and the historical marketing data to obtain the transaction scale data;

[0121] Determine recommended products and push personnel based on transaction size data, push recommended products to customers and connect the push personnel's signal source to the customer's account.

[0122] In summary, the computer-readable storage medium of the embodiment of the present invention first generates the similarities of each product based on the data of each financial product and the customer's ideal product data generated by the customer's preferred product information, and then generates each transaction scale data based on the similarities of each product and the marketing forecast data of each marketer to determine the recommended products and push personnel and push them to the customer, which can meet the real demands of the customer and improve the accuracy of product push.

[0123] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0124] Those skilled in the art may also understand that the various illustrative logical blocks, units, and steps listed in the embodiments of the present invention may be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly demonstrate the interchangeability of hardware and software, the various illustrative components, units, and steps described above have generally described their functions. Whether such functions are implemented by hardware or software depends on the specific application and the design requirements of the entire system. Those skilled in the art may use various methods to implement the described functions for each specific application, but such implementation should not be understood as exceeding the scope of protection of the embodiments of the present invention.

[0125] The various illustrative logic blocks, or units, or devices described in the embodiments of the present invention can be implemented or operated by a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array or other programmable logic device, a discrete gate or transistor logic, a discrete hardware component, or any combination of the above. The general-purpose processor can be a microprocessor, and optionally, the general-purpose processor can also be any conventional processor, controller, microcontroller or state machine. The processor can also be implemented by a combination of computing devices, such as a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a digital signal processor core, or any other similar configuration.

[0126] The steps of the method or algorithm described in the embodiments of the present invention can be directly embedded in hardware, a software module executed by a processor, or a combination of the two. The software module can be stored in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, a CD-ROM, or other storage media of any form in the art. Exemplarily, the storage medium can be connected to the processor so that the processor can read information from the storage medium and can write information to the storage medium. Optionally, the storage medium can also be integrated into the processor. The processor and the storage medium can be arranged in an ASIC, and the ASIC can be arranged in a user terminal. Optionally, the processor and the storage medium can also be arranged in different components in the user terminal.

[0127] In one or more exemplary designs, the above functions described in the embodiments of the present invention can be implemented in hardware, software, firmware or any combination of the three. If implemented in software, these functions can be stored on a computer-readable medium, or transmitted in the form of one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media that facilitate the transfer of computer programs from one place to another. The storage medium can be any available medium that can be accessed by any general or special computer. For example, such computer-readable media can include but are not limited to RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage devices, or any other medium that can be used to carry or store program codes in the form of instructions or data structures and other forms that can be read by general or special computers, or general or special processors. In addition, any connection can be appropriately defined as a computer-readable medium, for example, if the software is transmitted from a website site, server or other remote resource through a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL) or wirelessly, such as infrared, wireless and microwave, it is also included in the defined computer-readable medium. The disk and disc include compact disk, laser disk, optical disk, DVD, floppy disk and blue-ray disk. Disks usually copy data magnetically, while discs usually copy data optically with lasers. The above combination can also be included in computer readable media.

Claims

1. A method for pushing financial products, characterized in that: include: Acquire customer preferred product information, and generate customer ideal product data based on the customer preferred product information; Generating product similarity of each financial product according to the customer's ideal product data and each financial product data; Input the similarity of each product and the marketing forecast data of each marketing personnel into the transaction scale model created based on the historical similarity of the product and the historical marketing data to obtain the transaction scale data; Determine the recommended product and push personnel according to each transaction scale data, push the recommended product to the customer and cut the signal source of the push personnel into the customer's account; The product similarity of each financial product is determined by the following formula: Among them, A i is the product similarity of the ith financial product, ρ i is the Euclidean distance of the i-th product; According to the similarity of each product and the marketing forecast data of each marketer, the transaction scale model created based on the historical similarity of the product and the historical marketing data is input to obtain the transaction scale data, including: The transaction size data corresponding to the i-th financial product recommended by the k-th marketer is obtained according to the following formula: X ki =α0+α1A i +α2B k ; Among them, X ki The transaction scale data corresponding to the recommendation of the i-th financial product for the k-th marketing personnel, α0 is the first scale model parameter, α1 is the second scale model parameter, α2 is the third scale model parameter, A i is the product similarity of the ith financial product, B k is the marketing prediction data corresponding to the k-th marketing personnel; The Euclidean distance of each product is determined based on the customer's ideal product data and each financial product. Specifically, the Euclidean distance of each product is determined by the following formula: Among them, ρ i is the Euclidean distance of the ith product, x i is the risk data of the ith financial product, y i is the income data of the ith financial product, z i is the term data of the ith financial product, x′ is the risk data of the customer’s ideal product, x′ is the dimensionless grade data, with a value of 1-5, y′ is the annualized rate of return of the customer’s ideal product, in %, and z′ is the term data of the customer’s ideal product, in days.

2. The financial product push method according to claim 1, characterized in that: Recommended products and push personnel are determined based on transaction scale data, including: The financial product corresponding to the maximum value of each transaction scale data is determined as the recommended product, and the marketing personnel corresponding to the maximum value of each transaction scale data is determined as the push personnel.

3. The financial product push method according to claim 1, characterized in that: Also includes: Obtain transaction behavior data and customer facial data corresponding to each marketing personnel; The transaction behavior data and customer face data corresponding to each marketing personnel are input into the marketing prediction model created based on the transaction behavior training data and the customer face training data to obtain the marketing prediction data of each marketing personnel.

4. The financial product push method according to claim 3, characterized in that: The customer face data includes customer expression ratio data; The product push method also includes: Acquire a customer's historical video, and acquire a customer's facial expression image based on the customer's historical video; Using a convolutional neural network to extract features from the customer expression image to obtain the probability of each customer expression; Determine the customer expression percentage data based on the expression probability of each customer.

5. The financial product push method according to claim 4, characterized in that: The customer face data also includes customer face facing data; The product push method also includes: Acquire a customer's facial image based on the customer's historical video; Input the customer's face image into the face angle recognition model created based on the face training image to obtain the customer's face angle data; Determine the customer's face facing data based on the customer's face angle data.

6. The financial product push method according to claim 4, characterized in that: Creating a deal size model based on product historical similarity and historical marketing data includes: Perform the following iterative processing: Determining scale forecast data based on the product historical similarity, the historical marketing data and scale model parameters; Determine a scale loss function according to the scale prediction data and the corresponding scale actual data; When the scale loss function is less than a preset scale loss function threshold, a transaction scale model is created according to the scale model parameters; otherwise, the scale model parameters are updated according to the scale loss function, and the iterative process is continued.

7. A financial product push device, characterized in that: include: An ideal product data generation module is used to obtain customer preferred product information and generate customer ideal product data based on the customer preferred product information; A product similarity module, used to generate product similarity of each financial product based on the customer's ideal product data and each financial product data; The transaction scale data module is used to input the similarity of each product and the marketing forecast data of each marketing personnel into the transaction scale model created based on the historical similarity of the product and the historical marketing data to obtain the transaction scale data; A push module is used to determine recommended products and push personnel according to each transaction scale data, push the recommended products to the customer and switch the signal source of the push personnel into the customer's account; The product similarity module is specifically used to determine the product similarity of each financial product through the following formula: Among them, A i is the product similarity of the ith financial product, ρ i is the Euclidean distance of the i-th product; According to the similarity of each product and the marketing forecast data of each marketer, the transaction scale model created based on the historical similarity of the product and the historical marketing data is input to obtain the transaction scale data, including: The transaction size data corresponding to the i-th financial product recommended by the k-th marketer is obtained according to the following formula: X ki =α0+α1A i +α2B k ; Among them, X ki The transaction scale data corresponding to the recommendation of the i-th financial product for the k-th marketing personnel, α0 is the first scale model parameter, α1 is the second scale model parameter, α2 is the third scale model parameter, A i is the product similarity of the ith financial product, B k is the marketing prediction data corresponding to the k-th marketing personnel; The Euclidean distance of each product is determined based on the customer's ideal product data and each financial product. Specifically, the Euclidean distance of each product is determined by the following formula: Among them, ρ i is the Euclidean distance of the ith product, x i is the risk data of the ith financial product, y i is the income data of the ith financial product, z i is the term data of the ith financial product, x′ is the risk data of the customer’s ideal product, x′ is the dimensionless grade data, with a value of 1-5, y′ is the annualized rate of return of the customer’s ideal product, in %, and z′ is the term data of the customer’s ideal product, in days.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the financial product push method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the financial product push method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the financial product push method described in any one of claims 1 to 6 are implemented.

Images