Potential customer prediction method and device, electronic equipment and program product

Abstract

The invention discloses a potential customer prediction method and device, electronic equipment and a program product, and relates to the technical field of big data, and the method comprises the steps: dividing customers according to transaction information in a first time period, and obtaining a first type of customers and a second type of customers; obtaining first transaction information of the first type of customers and second transaction information of the second type of customers from the transaction information; performing feature extraction on the first transaction information to obtain first feature information, and performing feature extraction on the second transaction information to obtain second feature information; and predicting potential customers in the first type of customers according to the first feature information, predicting potential customers in the second type of customers according to the second feature information, and pushing product information to the potential customers. Through the method and the device, the problem that accurate pushing of product information is affected due to the fact that sample information corresponding to different types of transaction information of customers is unbalanced when a machine learning model is adopted to identify potential customers purchasing financial products in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of big data technology, and in particular to a method, device, electronic device and program product for predicting potential customers. Background Art

[0002] In the bank's personal customer system, although some high-quality customers enjoy exclusive services from financial managers, it cannot be ignored that the customer group that is not managed by financial managers is even larger and contains huge business potential. This part of customers is often easily overlooked due to the lack of direct personalized services, but their large number has an undeniable impact on the bank's customer base, business growth and even market competitiveness. At the same time, the discovery of high-potential customers is a technical means often used by financial institutions in the process of product promotion.

[0003] At present, the identification of high-potential customers of financial products mainly relies on single-dimensional indicators such as the scale of financial products held by customers, investment years, and investment behavior. The single identification dimension leads to low accuracy in the identification results of potential customers, and reduces the accuracy of product information push by financial institutions to potential customers.

[0004] When using machine learning models to identify potential customers who purchase financial products in related technologies, the output of the model is inaccurate due to the imbalance of sample information corresponding to different categories of customer transaction information, which affects the accurate push of product information. No effective solution has been proposed so far. Summary of the invention

[0005] The main purpose of this application is to provide a potential customer prediction method, device, electronic device and program product to solve the problem in the related technology that when a machine learning model is used to identify potential customers who purchase financial products, the sample information corresponding to different categories of customer transaction information is unbalanced, resulting in inaccurate model output results, thereby affecting the accurate push of product information.

[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for predicting potential customers is provided, the method comprising: dividing customers according to transaction information within a first time period to obtain a first category of customers and a second category of customers, wherein the first category of customers are customers who have not traded a target financial product within the first time period, and the second category of customers are customers who have traded a target financial product within the first time period; obtaining first transaction information of the first category of customers and second transaction information of the second category of customers from the transaction information; performing feature extraction on the first transaction information to obtain first feature information, and performing feature extraction on the second transaction information to obtain second feature information; inputting the first feature information into a first prediction model to predict potential customers among the first category of customers, inputting the second feature information into a second prediction model to predict potential customers among the second category of customers, and pushing product information of the target financial product to potential customers among the first category of customers and potential customers among the second category of customers.

[0007] Furthermore, feature extraction is performed on the first transaction information to obtain first feature information, including: statistics are performed on first-category variables in the first transaction information according to different time dimensions to obtain first-category data, wherein the first-category variables represent variables represented by a single numerical value; second-category variables are calculated using an unsupervised binning algorithm to obtain second-category data, wherein the second-category variables represent variables represented by numerical values ​​within an interval; multiple variables in the first transaction information are calculated to obtain third-category data, wherein the multiple variables include: the first-category variables and the second-category variables; and first feature information corresponding to the first transaction information is determined based on the first-category data, the second-category data, and the third-category data.

[0008] Furthermore, before inputting the first characteristic information into the first prediction model to predict potential customers among the first category of customers, and inputting the second characteristic information into the second prediction model to predict potential customers among the second category of customers, the method also includes: collecting historical transaction information of historical customers in the financial institution; constructing a first sample set and a second sample set based on the historical transaction information; wherein the times at which the historical customers in the first sample set and the second sample set purchased financial products are different; performing feature extraction on the first sample set to obtain third characteristic information, and performing feature extraction on the second sample set to obtain fourth characteristic information; performing data preprocessing on the third characteristic information to obtain processed third characteristic information, and performing data preprocessing on the fourth characteristic information to obtain processed fourth characteristic information; using the processed third characteristic information to train the preset model to obtain the first prediction model, and using the processed fourth characteristic information to train the preset model to obtain the second prediction model.

[0009] Further, constructing the first sample set and the second sample set based on the historical transaction information includes: determining a first positive sample based on the historical transaction information belonging to the third category of customers, determining a first negative sample based on the historical transaction information belonging to the fourth category of customers, and constructing the first sample set based on the first positive sample and the first negative sample, wherein the third category of customers are customers who have not traded the target financial product in the second time period and have traded the target financial product in the third time period, and the fourth category of customers are customers who have not traded the target financial product in the second time period and have not traded the target financial product in the third time period, and the third time period is later than the second time period, and the number of the first positive samples is The ratio between the number of the second positive samples and the number of the first negative samples is a first preset ratio; the second positive sample is determined based on the historical transaction information belonging to the fifth category of customers, the second negative sample is determined based on the historical transaction information belonging to the sixth category of customers, and the second sample set is constructed based on the second positive sample and the second negative sample, wherein the fifth category of customers are customers who have traded the target financial product within the second time period and the target financial product within the third time period, and the sixth category of customers are customers who have traded the target financial product within the second time period and the target financial product within the third time period, and the ratio between the number of the second positive samples and the number of the second negative samples is a second preset ratio.

[0010] Furthermore, the processed third feature information is used to train the preset model to obtain the first prediction model, including: dividing the processed third feature information into a first training set and a first test set; respectively using the first training set to train N first preset models to obtain N third models, and the output results of the N first preset models; respectively using the first test set to test the N third models to obtain the output results of the N third models, and constructing a second test set based on the output results of the N third models; constructing a second training set based on the output results of the N first preset models and the first training set; using the second training set to train the second preset model, and using the second test set to optimize the trained second preset model to obtain the first prediction model.

[0011] Furthermore, data preprocessing is performed on the third characteristic information to obtain processed third characteristic information, including: determining missing information in the third characteristic information, filling the missing information in the third characteristic information according to the data type of the missing information, to obtain fifth characteristic information, wherein the data type of the missing information includes at least one of the following: transaction behavior type, time type, numerical type; deleting a preset character string in the fifth characteristic information to obtain sixth characteristic information, wherein the preset character string includes at least: a character string corresponding to repeated information, a character string corresponding to non-semantic information; based on the target indicator value, screening key features in the sixth characteristic information to obtain the processed third characteristic information, wherein the calculation method of the target indicator value includes at least: evidence weight, information value, random forest algorithm.

[0012] Furthermore, the first transaction information of the first category of customers includes at least the following information: basic information of the customer, asset information, liability information, transaction behavior information, and client behavior data.

[0013] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a potential customer prediction device is provided, which includes: a classification unit, which is used to classify customers according to transaction information within a first time period to obtain a first category of customers and a second category of customers, wherein the first category of customers are customers who have not traded the target financial product within the first time period, and the second category of customers are customers who have traded the target financial product within the first time period; an acquisition unit, which is used to obtain first transaction information of the first category of customers and second transaction information of the second category of customers from the transaction information; a first extraction unit, which is used to perform feature extraction on the first transaction information to obtain first feature information, and perform feature extraction on the second transaction information to obtain second feature information; a push unit, which is used to input the first feature information into a first prediction model to predict potential customers in the first category of customers, input the second feature information into a second prediction model to predict potential customers in the second category of customers, and push product information of the target financial product to potential customers in the first category of customers and potential customers in the second category of customers.

[0014] Furthermore, the first extraction unit includes: a statistical subunit, which is used to perform statistics on the first type of variables in the first transaction information according to different time dimensions to obtain the first type of data, wherein the first type of variables represent variables represented by a single numerical value; a first calculation subunit, which is used to calculate the second type of variables in the first transaction information using an unsupervised binning algorithm to obtain the second type of data, wherein the second type of variables represent variables represented by numerical values ​​within an interval range; a second calculation subunit, which is used to calculate multiple variables in the first transaction information to obtain the third type of data, wherein the multiple variables include: the first type of variables and the second type of variables; a first determination subunit, which is used to determine the first feature information corresponding to the first transaction information based on the first type of data, the second type of data and the third type of data.

[0015] Furthermore, the device also includes: a collection unit, which is used to collect historical transaction information of historical customers in a financial institution before inputting the first feature information into a first prediction model to predict potential customers among the first category of customers and inputting the second feature information into a second prediction model to predict potential customers among the second category of customers; a construction unit, which is used to construct a first sample set and a second sample set based on the historical transaction information; wherein the time when the historical customers in the first sample set and the second sample set purchased financial products is different; a second extraction unit, which is used to perform feature extraction on the first sample set to obtain third feature information, and perform feature extraction on the second sample set to obtain fourth feature information; a processing unit, which is used to perform data preprocessing on the third feature information to obtain processed third feature information, and perform data preprocessing on the fourth feature information to obtain processed fourth feature information; a training unit, which is used to train a preset model using the processed third feature information to obtain the first prediction model, and to train the preset model using the processed fourth feature information to obtain the second prediction model.

[0016] Further, the construction unit includes: a second determination subunit, used to determine a first positive sample based on the historical transaction information belonging to the third category of customers, determine a first negative sample based on the historical transaction information belonging to the fourth category of customers, and construct the first sample set based on the first positive sample and the first negative sample, wherein the third category of customers are customers who have not traded the target financial product in the second time period and have traded the target financial product in the third time period, the fourth category of customers are customers who have not traded the target financial product in the second time period and have not traded the target financial product in the third time period, the third time period is later than the second time period, and the number of the first positive samples and the number of the first negative samples are The ratio between the quantities is a first preset ratio; the third determination subunit is used to determine the second positive sample based on the historical transaction information belonging to the fifth category of customers, determine the second negative sample based on the historical transaction information belonging to the sixth category of customers, and construct the second sample set based on the second positive sample and the second negative sample, wherein the fifth category of customers are customers who have traded the target financial product within the second time period and the target financial product within the third time period, and the sixth category of customers are customers who have traded the target financial product within the second time period and the target financial product within the third time period, and the ratio between the number of the second positive samples and the number of the second negative samples is a second preset ratio.

[0017] Furthermore, the training unit includes: a division subunit, used to divide the processed third feature information into a first training set and a first test set; a first training subunit, used to respectively use the first training set to train N first preset models, to obtain N third models, and output results of the N first preset models; a first construction subunit, used to respectively use the first test set to test the N third models, to obtain the output results of the N third models, and to construct a second test set based on the output results of the N third models; a second construction subunit, used to construct a second training set based on the output results of the N first preset models and the first training set; a second training subunit, used to train the second preset model using the second training set, and to optimize the trained second preset model using the second test set to obtain the first prediction model.

[0018] Furthermore, the processing unit includes: a filling subunit, used to determine the missing information in the third characteristic information, and fill the missing information in the third characteristic information according to the data type of the missing information to obtain the fifth characteristic information, wherein the data type of the missing information includes at least one of the following: transaction behavior type, time type, numerical type; a deleting subunit, used to delete the preset character string in the fifth characteristic information to obtain the sixth characteristic information, wherein the preset character string includes at least: a character string corresponding to repeated information, and a character string corresponding to non-semantic information; a screening subunit, used to screen key features in the sixth characteristic information based on the target indicator value to obtain the processed third characteristic information, wherein the calculation method of the target indicator value includes at least: evidence weight, information value, and random forest algorithm.

[0019] Furthermore, the first transaction information of the first category of customers includes at least the following information: basic information of the customer, asset information, liability information, transaction behavior information, and client behavior data.

[0020] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a computer program product is provided, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for predicting potential customers, and which, when executed by a processor, implements the steps of the methods for predicting potential customers described in each embodiment of the present application.

[0021] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes stored computer instructions, wherein when the computer instructions are executed by a processor, any one of the above-mentioned potential customer prediction methods is implemented.

[0022] In order to achieve the above-mentioned purpose, according to one aspect of the present application, an electronic device is provided, including one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement any of the above-mentioned potential customer prediction methods.

[0023] In an embodiment of the present application, customers are divided according to transaction information within a first time period to obtain a first category of customers and a second category of customers, wherein the first category of customers are customers who have not traded a target financial product within the first time period, and the second category of customers are customers who have traded a target financial product within the first time period; first transaction information of the first category of customers and second transaction information of the second category of customers are obtained from the transaction information; feature extraction is performed on the first transaction information to obtain first feature information, and feature extraction is performed on the second transaction information to obtain second feature information; the first feature information is input into a first prediction model to predict potential customers among the first category of customers, the second feature information is input into a second prediction model to predict potential customers among the second category of customers, and product information of the target financial product is pushed to the potential customers among the first category of customers and the potential customers among the second category of customers, thereby solving the technical problem that when a machine learning model is used to identify potential customers who purchase financial products, the model output result is inaccurate due to the imbalance of sample information corresponding to transaction information of different categories of customers, thereby affecting the accurate push of product information.

[0024] By extracting features from customers' transaction data and using the extracted feature information to train prediction models, it is possible to predict the likelihood of each customer becoming a potential customer of the target financial product, allowing banks to accurately and effectively identify customers with high purchasing potential. At the same time, by classifying non-client manager accounts into new and old customers based on transaction behavior, financial institutions can narrow the scope of recommended customers, reduce the cost and waste of resources of blindly reaching a large number of non-target customers, and push different product information to different types of customers, thereby improving the accuracy of financial product information push and customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0026] Figure 1 It is a hardware structure block diagram of a computer terminal (or mobile device) for implementing a potential customer prediction method provided in accordance with the first embodiment of the present application;

[0027] Figure 2 is a flowchart of an optional potential customer prediction method provided according to Embodiment 1 of the present application;

[0028] Figure 3 It is a schematic diagram of an optional process of training two prediction models according to the first embodiment of the present application;

[0029] Figure 4is a flowchart of optional customer classification provided according to Embodiment 1 of the present application;

[0030] Figure 5 It is a schematic diagram of an optional observation period and a performance period in the process of predicting potential customers provided in the first embodiment of the present application;

[0031] Figure 6 It is a schematic diagram of an optional process of integrated learning based on a primary learner and a secondary learner provided in accordance with the first embodiment of the present application;

[0032] Figure 7 is a schematic diagram of a potential customer prediction device provided according to Embodiment 2 of the present application;

[0033] Figure 8 It is a schematic diagram of an electronic device for predicting potential customers provided according to Embodiment 5 of the present application. DETAILED DESCRIPTION

[0034] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0035] It should be noted that the methods and devices determined by the processing methods, devices, storage media and electronic devices of the present application documents can be used in the field of financial technology to improve the accuracy of pushed product information in the process of pushing financial products to potential customers who purchase financial products. They can also be used in any field other than the field of financial technology. The application fields of the methods and devices of the processing methods, devices, storage media and electronic devices of the present application documents are not limited.

[0036] It should be noted that the collected information (including but not limited to user device information, user personal information, collected data, used data, generated data, processed data, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, collected information, used information, generated information, processed information, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with the relevant laws, regulations and standards of relevant countries and regions, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation entrances are provided for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or institutions to provide users with corresponding operation entrances for users to choose to agree or refuse the results of automated decision-making; if the user chooses to refuse, the expert decision-making process will be entered.

[0037] Example 1

[0038] According to an embodiment of the present application, a method embodiment for predicting potential customers is also provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0039] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computer terminal (or mobile device) for implementing a method for predicting potential customers. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more (102a, 102 potential customer prediction, ..., 102n are used to illustrate) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the USB bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0040] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0041] The memory 104 can be used to store software programs and modules of application software, such as program instructions / data storage devices corresponding to the method for predicting potential customers in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the method for predicting potential customers described above is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0042] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0043] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0044] Under the above operating environment, this application provides Figure 2 The forecasting method for potential customers is shown. Figure 2 It is a flowchart of the method for predicting potential customers according to Example 1 of the present application.

[0045] Step S101, classifying customers according to transaction information in a first time period to obtain first-category customers and second-category customers, wherein the first-category customers are customers who have not traded the target financial product in the first time period, and the second-category customers are customers who have traded the target financial product in the first time period.

[0046] In the first embodiment, the first time period can be defined as the past 6 months, that is, the time window of the last 180 days. In this time window, the financial institution analyzes the non-client manager-managed customers based on the transaction information of the financial products, identifies the customer's transaction behavior, and thus divides the non-client manager-managed customers into the first category of customers and the second category of customers. Exemplarily, the target financial product can be a fund product.

[0047] The first type of customer refers to customers who have not purchased funds (target financial products) in the first defined period of time (the past 6 months) and are not managed by account managers, i.e., new customers, and can also be called potential new purchasers. The second type of customer refers to customers who have purchased funds (i.e., the above-mentioned target financial products) at least once in the first defined period of time (the past 6 months) and are not managed by account managers, i.e., old customers.

[0048] By segmenting customers based on transaction information, financial institutions can accurately implement differentiated promotion strategies for different types of customer groups, such as introducing and promoting products to new customers, and improving the accuracy of product information push by increasing transaction frequency and investment amount for old customers. Through this precise classification, finance can allocate resources more effectively and improve customer satisfaction.

[0049] Step S102, obtaining first transaction information of first-category customers and second transaction information of second-category customers from the transaction information.

[0050] In the first embodiment, the first transaction information refers to the transaction information of the first type of customers during the observation period (i.e., data from the past 12 months), for example, the basic information of the first type of customers (such as age, occupation), other financial products held in the bank (such as deposit balance, loan balance), fund transaction behavior (such as frequency of use of mobile banking, number of transactions), transaction behavior or customer characteristics associated with transactions of multiple financial products, etc.

[0051] The second transaction information of the second type of customers refers to the fund transaction records and other transaction information related to fund transactions during the observation period (i.e., the data of the past 12 months). The method of obtaining the second transaction information of the second type of customers is similar to the method of obtaining the first transaction information of the first type of customers.

[0052] By obtaining the first transaction information, it is possible to identify customers who have never purchased a fund but are likely to become new customers in the future. By obtaining the second transaction information, it is possible to analyze the existing fund transaction information and, in combination with other transaction behaviors, identify old customers who have the potential to increase the frequency or amount of fund investment.

[0053] By obtaining and analyzing the transaction information of the first and second types of customers in the above manner, data mining and machine learning technologies can be used more effectively to accurately classify and recommend potential customers, while reducing promotion costs and improving the accuracy of financial product information push and customer satisfaction.

[0054] Step S103: extracting features from the first transaction information to obtain first feature information, and extracting features from the second transaction information to obtain second feature information.

[0055] In the first embodiment, feature extraction is performed on the first transaction information, aiming to extract features that can reflect the willingness of new customers to purchase financial products from these first transaction data. For example, by focusing on the customer's age, asset status, frequency of mobile banking use, whether they have browsed financial product pages, whether they have transaction records of other financial products (such as loans, financial management), etc., the first feature information is formed, and the first feature information is used to train the model to predict potential customers with a high probability of becoming new customers.

[0056] Feature extraction of the second transaction information will conduct an in-depth analysis of the second transaction information, for example, analyzing the customer's transaction frequency, distribution of transaction amounts, transaction time points, types of financial products held, financial product page browsing behavior on mobile banking, customer service star ratings, etc., to reflect old customers' interest in financial products and investment behavior patterns, forming the second feature information, which is used to train the model to predict potential customers among old customers who have the potential to increase fund investment frequency or increase investment amount.

[0057] By converting raw transaction data into targeted feature information, the predictive model can more accurately learn and predict customer behavior patterns, thereby improving the accuracy and effectiveness of financial product information push.

[0058] Step S104, input the first characteristic information into the first prediction model to predict potential customers among the first category of customers, input the second characteristic information into the second prediction model to predict potential customers among the second category of customers, and push product information of the target financial product to the potential customers among the first category of customers and the potential customers among the second category of customers.

[0059] In the first embodiment, the first characteristic information is input into a trained first prediction model (also referred to as a new customer acquisition model), and based on the first characteristic information, the fund purchase probability of each first-category customer is calculated, thereby predicting potential new customers among the first-category customers.

[0060] The second feature information is input into the trained second prediction model (also called quality improvement model) to predict the second type of customers who have the potential to increase their fund investment and become old customers with higher value.

[0061] After the first prediction model and the second prediction model identify potential new customers in the first category and potential old customers in the second category, targeted promotion strategies can be adopted. For example, financial product information can be pushed to the above potential customers, including detailed descriptions of the products, purchase processes, possible benefits and risks, promotional activities, etc., in order to stimulate their investment interest and increase fund sales. The accuracy and personalization of the pushed information usually directly affect customers' cognition and purchase decisions of fund products. Therefore, the push content can be designed according to the model prediction results to meet the needs of different potential customers.

[0062] Through the above process, financial institutions can effectively utilize machine learning models to accurately identify potential buyers and potential quality-upgrading customers of fund products based on the characteristics of customer behavior and transaction data, thereby optimizing promotion strategies and improving the accuracy of financial product information push and customer satisfaction.

[0063] Optionally, in the potential customer prediction method provided in Example 1 of the present application, feature extraction is performed on the first transaction information to obtain first feature information, including: statistics are performed on the first category of variables in the first transaction information according to different time dimensions to obtain first category data, wherein the first category of variables represent variables represented by a single numerical value; an unsupervised binning algorithm is used to calculate the second category of variables in the first transaction information to obtain second category data, wherein the second category of variables represent variables represented by numerical values ​​within an interval range; multiple variables in the first transaction information are calculated to obtain third category data, wherein the multiple variables include: first category variables, second category variables; and the first feature information corresponding to the first transaction information is determined based on the first category data, the second category data, and the third category data.

[0064] In the first embodiment of the present invention, the first type of variables refers to variables that can be directly represented by a single value, such as transaction amount, wage payment, number of financial product browsing times, number of mobile banking logins in the past year, etc. These variables are counted according to different time dimensions (such as monthly, quarterly, and annual) to obtain the first type of data. For example, for the number of mobile banking logins, the bank may count the number of customer logins in the past 1 month, the past 3 months, and the past 12 months to reflect the customer's recent and long-term mobile banking usage habits. The statistics of these data are intended to capture the dynamic behavior and financial status of customers and provide stable and interpretable features for the model.

[0065] The second type of variables refers to variables that need to be represented by values ​​within an interval, such as customer age, asset information, etc. For this type of variables, unsupervised binning algorithms can be used for discretization. Unsupervised binning algorithms can convert continuous values ​​into bins or intervals, with each interval representing a category. For example, customer age can be binned into "18-25 years old", "26-35 years old", "36-45 years old", etc. This can reduce the impact of outliers and enable the model to better handle nonlinear relationships.

[0066] The third type of data is obtained by calculating the combination of the first and second types of variables. For example, you can calculate the combined features of customer age and deposit balance, or the correlation between the number of logins and browsing behavior in recent months. These combined features can usually capture the complex relationship between customer behavior and financial status, providing richer information for the model.

[0067] Finally, the first feature information mentioned above is obtained by combining the first, second and third types of data. By extracting feature information from the original transaction data or creating new feature information based on the original transaction data, it is possible to capture relationships and patterns that are not directly reflected in the original data, helping the model to better understand and predict customer behavior. At the same time, it can also increase the generalization ability of the model and reduce overfitting of the training data set, thereby helping the prediction model to output more accurate prediction results, further achieving the effect of improving the accuracy of financial product information push and customer satisfaction.

[0068] Optionally, in the potential customer prediction method provided in Example 1 of the present application, before inputting the first feature information into the first prediction model to predict potential customers among the first category of customers, and inputting the second feature information into the second prediction model to predict potential customers among the second category of customers, the above method also includes: collecting historical transaction information of historical customers in the financial institution; constructing a first sample set and a second sample set based on the historical transaction information; wherein the time when the historical customers in the first sample set and the second sample set purchase financial products are different; performing feature extraction on the first sample set to obtain third feature information, and performing feature extraction on the second sample set to obtain fourth feature information; performing data preprocessing on the third feature information to obtain processed third feature information, and performing data preprocessing on the fourth feature information to obtain processed fourth feature information; using the processed third feature information to train the preset model to obtain the first prediction model, and using the processed fourth feature information to train the preset model to obtain the second prediction model.

[0069] In the first embodiment of the present invention, the financial institution will collect historical transaction information of its historical customers on various transaction activities, including but not limited to fund purchase records, deposit operations, loan information, mobile banking usage, etc. According to the time attribute of the historical transaction information, the first sample set and the second sample set are constructed respectively. The first sample set contains the customer data who did not purchase the fund during the observation period (such as the past 12 months) but purchased the fund in the next time window (such as 1 month after the observation period). These customers will be defined as positive samples in the first sample set. The second sample set contains the customer data who purchased the fund during the observation period and continued to purchase the fund within 1 month after the observation period. These customers will be defined as positive samples in the second sample set.

[0070] Then, feature extraction is performed on the first sample set to obtain third feature information, which may include the customer's basic information, asset-liability status, product holdings, fund transaction behavior, mobile banking usage habits, etc. Similarly, feature extraction is performed on the second sample set to obtain fourth feature information. The feature extraction process can adopt the above-mentioned feature extraction process for the first transaction information of the first type of customer, so as to more accurately describe the customer's behavior and attributes.

[0071] Secondly, the third characteristic information and the fourth characteristic information are subjected to corresponding data preprocessing to obtain the processed third characteristic information and the processed fourth characteristic information. The purpose of preprocessing is to clean the data, fill in the missing information, and convert the data format, thereby improving the prediction accuracy and stability of the model.

[0072] Finally, the preset model is trained using the processed third feature information to obtain the first prediction model, i.e., the new customer acquisition model. The new customer acquisition model is specifically used to predict potential new customers who will purchase financial products in the future. Similarly, the preset model is trained using the processed fourth feature information to obtain the second prediction model, i.e., the quality improvement model. The quality improvement model is used to predict old customers who have the potential to increase fund investment or increase their investment frequency among customers who have already purchased funds.

[0073] By collecting historical customer transaction information, we construct two sample sets separated by time, extract features from these two sample sets respectively, perform data preprocessing, and finally use the processed feature information to train two different prediction models (new customer acquisition model and quality improvement model). This enables financial institutions to accurately predict potential new customers and potential old customers who will purchase financial products based on machine learning models, thereby implementing more effective promotion strategies and improving the accuracy of financial product information push and customer satisfaction.

[0074] Optionally, in the method for predicting potential customers provided in the first embodiment of the present application, constructing a first sample set and a second sample set based on historical transaction information includes: determining a first positive sample based on historical transaction information belonging to a third category of customers, determining a first negative sample based on historical transaction information belonging to a fourth category of customers, and constructing the first sample set based on the first positive sample and the first negative sample, wherein the third category of customers are customers who have not traded the target financial product in the second time period and have traded the target financial product in the third time period, the fourth category of customers are customers who have not traded the target financial product in the second time period and have not traded the target financial product in the third time period, and the third time period is later than the second time period. time period, the ratio between the number of first positive samples and the number of first negative samples is a first preset ratio; the second positive sample is determined based on the historical transaction information of the fifth category of customers, the second negative sample is determined based on the historical transaction information of the sixth category of customers, and a second sample set is constructed based on the second positive sample and the second negative sample, wherein the fifth category of customers are customers who have traded the target financial product in the second time period and the target financial product in the third time period, the sixth category of customers are customers who have traded the target financial product in the second time period and the target financial product in the third time period, and the ratio between the number of second positive samples and the number of second negative samples is a second preset ratio.

[0075] In the first embodiment, the third type of customers refers to customers who have not traded the target financial product (e.g., fund) in the second time period (e.g., within 6 months), but have traded the fund in the subsequent third time period (e.g., within 1 month after 6 months). Then, the first positive sample in the first sample set is formed based on the historical transaction information of the third type of customers in the observation period (e.g., 12 months).

[0076] The fourth type of customers refers to customers who have not traded the target financial product (e.g., fund) in the second time period (e.g., within 6 months), but have not traded funds in the subsequent third time period (e.g., within 1 month after 6 months). Then, the first negative sample in the first sample set is formed based on the historical transaction information of the fourth type of customers in the observation period (e.g., 12 months).

[0077] The determined first positive sample and the first negative sample are combined into a first sample set for training the new customer acquisition model, i.e., the first prediction model mentioned above. It should be noted that when constructing the first sample set, attention should be paid to the ratio between the number of positive and negative samples, i.e., the first preset ratio. For example, through subsequent steps such as data preprocessing and feature extraction, the positive-negative sample ratio is adjusted from the original extremely unbalanced state to a more reasonable ratio, such as from 1:1500 to 1:500, so as to improve the efficiency of model training and the accuracy of prediction.

[0078] The fifth type of customers refers to customers who have traded the target financial product (e.g., fund) in the second time period (e.g., within 6 months), but traded the fund again in the subsequent third time period (e.g., within 1 month after 6 months). Then, the second positive sample in the second sample set is formed based on the historical transaction information of the fifth type of customers in the observation period (e.g., 12 months).

[0079] The sixth category of customers refers to customers who have traded the target financial product (e.g., fund) in the second time period (e.g., within 6 months) but have not traded funds in the subsequent third time period (e.g., within 1 month after 6 months). Then, the second negative sample in the second sample set is formed based on the historical transaction information of the sixth category of customers in the observation period (e.g., 12 months).

[0080] The determined second positive sample and the second negative sample are combined into a second sample set for training the quality improvement model, i.e., the second prediction model mentioned above. It should also be noted that the ratio between the number of positive and negative samples in the second sample set (and the second preset ratio mentioned above) needs to be reasonably adjusted, such as from 1:45 to 1:30, to ensure that the feature differences of positive and negative samples can be fully learned during the model training process, thereby improving the generalization ability and prediction accuracy of the model.

[0081] Through the above steps, the customer groups can be accurately divided, and the first sample set and the second sample set can be constructed to train the new customer acquisition model and the quality improvement model respectively. The two models are optimized for different customer groups (new customers and old customers) to predict new customers with potential fund purchases or old customers with potential for increased investment, thereby achieving the effect of narrowing the scope of potential customers, reducing promotion costs, and improving the accuracy of financial product information push and customer satisfaction. At the same time, by adjusting the positive and negative sample ratios in the first sample set and the second sample set, the model can be prevented from overfitting and the accuracy of the model prediction results can be improved.

[0082] Optionally, in the potential customer prediction method provided in Example 1 of the present application, the processed third feature information is used to train the preset model to obtain a first prediction model, including: dividing the processed third feature information into a first training set and a first test set; using the first training set to train N first preset models respectively to obtain N third models, and the output results of the N first preset models; using the first test set to test the N third models respectively to obtain the output results of the N third models, and constructing a second test set based on the output results of the N third models; constructing a second training set based on the output results of the N first preset models and the first training set; using the second training set to train the second preset model, and using the second test set to optimize the trained second preset model to obtain the first prediction model.

[0083] In the first embodiment of the present invention, the processed third feature information (feature information for the new model) is divided into a first training set and a first test set. The N first preset models are trained using the first training set to obtain N third models. The N first preset models here refer to multiple pre-selected machine learning models, such as LightGBM, random forest, and decision tree. The N third models can also be referred to as N primary learners.

[0084] Then, the N third models trained with the first training set are used to perform predictions on the first test set to obtain the output results of the N third models. The output results of the N third models for the first test set will be used to construct the second test set.

[0085] Secondly, based on the output results of the N first preset models and the first training set, a second training set is constructed. At this stage, the second training set not only contains the original first training set, but also adds the output results of the primary learner as additional features, as well as the label information corresponding to each feature. This combined feature will be used to train the second preset model (also known as the secondary learner). Next, the N third models are tested using the first test set to obtain the output results of the N third models.

[0086] Secondly, the second preset model (also called a meta-learner, or a secondary learner, the second preset model can be a logistic regression model) is trained using the second training set to obtain the trained second preset model. In this process, the second preset model will combine the output results of the N first preset models and the original feature information (i.e., the first training set mentioned above) to provide more accurate prediction results. Subsequently, the trained second preset model is tested and fine-tuned using the second test set to obtain the above-mentioned first prediction model (new user model) to evaluate the overall performance of the model and adjust the parameters to improve the prediction accuracy.

[0087] In an optional embodiment, the processed fourth feature information (feature information for the quality improvement model) is divided into a third training set and a third test set. The N third preset models are trained using the third training set to obtain N fourth models. The N third preset models here can also use models such as LightGBM, random forest and decision tree. The N fourth models can also be called N primary learners.

[0088] Then, the N fourth models are used to perform predictions on the third test set respectively to obtain the output results of the N fourth models. The output results of the N fourth models for the third test set will be used to construct the fourth test set.

[0089] Secondly, a third training set is constructed based on the output results of the N third preset models and the second training set. At this stage, the third training set not only contains the original third training set, but also adds the output results of the primary learner as additional features, as well as the label information corresponding to each feature. This combined feature will be used to train the fourth preset model (also referred to as a secondary learner). Next, the N fourth models are tested using the fourth test set to obtain the output results of the N fourth models.

[0090] Secondly, the fourth training set is used to train the fourth preset model (also called a meta-learner, or a secondary learner, the fourth preset model can be a logistic regression model) to obtain the trained fourth preset model. In this process, the fourth preset model will combine the output results of the N third preset models and the original feature information (ie, the third training set mentioned above) to provide more accurate prediction results. Subsequently, the trained fourth preset model is tested and fine-tuned using the fourth test set to obtain the above-mentioned second prediction model (quality improvement model) to evaluate the overall performance of the model and adjust the parameters to improve the prediction accuracy.

[0091] Through the above steps, an integrated and optimized first prediction model (new customer acquisition model) and second prediction model (quality improvement model) can be obtained. The prediction effect of this model is usually better than that of a single base model, especially in processing complex data and avoiding overfitting. It can output more accurate prediction results, which will help financial institutions implement more effective promotion strategies and improve the accuracy of financial product information push and customer satisfaction.

[0092] Optionally, in the potential customer prediction method provided in Example 1 of the present application, data preprocessing is performed on the third characteristic information to obtain processed third characteristic information, including: determining missing information in the third characteristic information, filling the missing information in the third characteristic information according to the data type of the missing information, to obtain fifth characteristic information, wherein the data type of the missing information includes at least one of the following: transaction behavior type, time type, numerical type; deleting a preset character string in the fifth characteristic information to obtain sixth characteristic information, wherein the preset character string includes at least: a character string corresponding to repeated information, a character string corresponding to non-semantic information; based on the target indicator value, key features are screened in the sixth characteristic information to obtain the processed third characteristic information, wherein the calculation method of the target indicator value includes at least: evidence weight, information value, random forest algorithm.

[0093] In the first embodiment, the missing data in the third characteristic information (i.e., the characteristic information extracted from historical customers for building a new customer acquisition model) is determined. Since the missing information may involve different data types (e.g., transaction behavior type (e.g., fund transaction behavior), time type (e.g., transaction date), and numerical type (e.g., asset balance)), different filling strategies can be adopted according to the data type of the missing information.

[0094] In an optional embodiment, for missing data belonging to the transaction behavior type, that is, data missing due to the non-occurrence of the behavior, it can be filled with 0 (or a specific mark) to indicate that the behavior does not exist. For example, the missing data caused by the customer not opening a certain service, not holding a certain financial product, or not having a behavior can be filled with 0. For missing data belonging to the time type, it can be filled with a special value (such as 999). For example, if the customer has not purchased a fund through mobile banking, if the missing data is the time of the last login to mobile banking, filling 999 means that the customer has not logged into mobile banking for a long time. For missing data belonging to the numerical type, it can be filled with statistical values ​​such as mean and median. For example, if the missing data is the customer's age, it can be filled according to the mean age of historical customers to maintain the distribution characteristics of the numerical feature. After the above-mentioned missing value filling process, the complete fifth feature information is obtained, which ensures the integrity of the feature information.

[0095] Then, the preset strings in the fifth feature information are deleted. These strings usually represent meaningless information, such as repeated field names, empty fields, or field values ​​that are irrelevant to the prediction target. The preset strings can be predefined to identify specific types of useless information. By eliminating meaningless information, a more streamlined sixth feature information can be obtained, which reduces the impact of noise data on model training and improves data quality and the efficiency of subsequent feature screening.

[0096] Finally, the features in the sixth feature information are screened based on the target index value to retain feature information that is of great value for predicting potential customers.

[0097] The calculation method of the target indicator value includes but is not limited to: Weight of Evidence (WOE): measures the strength of the relationship between the feature value and the positive target value, and is used for quantitative evaluation of the feature; Information Value (IV) is calculated based on WOE, reflecting the ability of the feature to distinguish between positive and negative samples. Features with high IV values ​​usually contribute more to prediction; Use feature importance scores during random forest model training to identify the features that have the greatest impact on the model prediction results.

[0098] In an optional embodiment, a first variable with an information value greater than 0.02 can be selected, and based on the first variable importance ranking result given by the random forest model, the top 80 ranked variables with importance indicators (e.g., feature_importance) greater than 0 can be screened to obtain the above-mentioned processed third feature information.

[0099] In an optional embodiment, according to the steps of performing data preprocessing on the third characteristic information to obtain the processed third characteristic information, the fourth characteristic information can be preprocessed to obtain the processed fourth characteristic information.

[0100] Through this series of preprocessing steps, we can ensure that the feature information used for model training is both complete and highly predictive, thereby improving the accuracy and stability of the model, reducing the risk of overfitting, and being able to more accurately predict potential new fund customers, thereby improving the accuracy of financial product information push and customer satisfaction.

[0101] Optionally, in the potential customer prediction method provided in Example 1 of the present application, the first transaction information of the first type of customer includes at least the following information: basic information of the customer, asset information, liability information, transaction behavior information, and client behavior data.

[0102] In the first embodiment, the basic information involves the personal attributes of the customer, including but not limited to customer service star rating, education level, marital status, occupation, customer age and region of origin. This basic information helps to understand the background of the customer, such as their possible investment risk preference, income level and financial stability. It should be noted that the step of collecting the basic information of the customer is performed after obtaining the authorization information of the customer.

[0103] Asset information reflects the client's asset holdings in banks or other financial institutions, including but not limited to deposit balances, basic factoring balances (may refer to funds, insurance, wealth management products, etc.), the amount of low, medium and high risk funds held by the client and the average daily assets in the past 3, 6 and 12 months, local and foreign currency wealth management, third-party deposits, account crude oil, credit cards, paper gold opening signs, personal consumption loans, personal housing loans, personal business loan opening signs, etc. Asset information is an important indicator for evaluating the client's investment potential and financial status.

[0104] Liability information includes loan balances, credit card overdraft balances, and other credit card liabilities. By analyzing a customer's liability structure, one can assess their financial health and debt level, which is critical to predicting their investment behavior and fund purchase potential.

[0105] Transaction behavior information records the details of the customer's transactions with the bank, including the purchase of financial products through mobile banking, the number and amount of debit card transactions, the amount of credit card repayments, the number and amount of financial products purchased on the Internet, etc. Transaction behavior information provides direct evidence of the customer's transaction habits and helps predict their interest and ability to purchase funds.

[0106] Client behavior data mainly comes from the customer's mobile banking behavior records, including the time since the last login to mobile banking, whether there is a login mark for mobile banking in the past 30 days, the number and time of browsing my account through mobile banking in the past month, the number of browsing the task center in the past month, and the number of logins to mobile banking in the past 1, 3, and 6 months, etc. These behavior data reflect the customer's activity and interest in banking services, and are of great value in understanding the customer's investment intentions and predicting their fund purchase behavior.

[0107] By collecting and analyzing the above multi-dimensional first transaction information, a detailed customer portrait can be constructed, which not only helps the model understand the customer's basic situation and financial status, but also captures the customer's trading habits and preferences for specific financial products, thereby improving the model's accuracy and effectiveness in predicting new fund customers. Ultimately, this series of information will be used to build and train a new customer acquisition model to identify new customers with high potential for fund purchases, thereby achieving the effect of improving the accuracy of financial product information push and customer satisfaction.

[0108] Optionally, in this embodiment 1, Figure 3 Figure 1 is a flowchart of training two prediction models. First, the personal banking customers who are not directly managed by the bank's account managers are stratified according to whether they have purchased funds in the past 6 months, and two groups of "old customers" and "new customers" are obtained.

[0109] The "old customers" group further enters the "quality improvement model" process, the goal of this model is to identify potential old customers who may purchase funds again in the next month. Before training the quality improvement model, the old customer data set is first screened to narrow the sample range and address the impact of unbalanced sample data to improve model efficiency and prediction accuracy. Next, feature construction is performed on the screened samples, including feature extraction, processing missing values ​​and outliers, and using information value (IV), random forest feature importance and other techniques to screen key features. Finally, the predictions of multiple preset models (such as LightGBM, random forest, decision tree) are integrated and trained with logistic regression as a meta-learner to improve the accuracy of the prediction.

[0110] The "new customer" group enters the "new customer acquisition model" training process, the goal is to identify which new customers are likely to buy funds for the first time in the next month. The three steps of sample range screening, feature construction, processing and screening, and integrated learning are similar to the process of the quality improvement model, but are targeted at new customer groups, with the goal of more accurately predicting the possibility of new customers buying funds.

[0111] Optionally, in this embodiment 1, Figure 4 This is a flow chart of customer classification. First, all customers in the financial institution who are not directly managed by a wealth manager are identified, and then classified according to whether they have purchased funds in the past 180 days. If a customer has purchased a fund at least once in the past 180 days, he or she is classified as an "old customer". These customers have a certain level of knowledge and investment experience in fund products and are potential targets for banks to increase their investment amount and frequency. On the contrary, if a customer has no record of purchasing a fund in the past 180 days, he or she is classified as a "new customer". This may include customers who are considering investing in a fund for the first time, or customers who have not invested in a fund in the past period of time.

[0112] Optionally, in this embodiment 1, Figure 5 It is a schematic diagram of the observation period and performance period in the process of predicting potential customers. It clearly depicts the concept of time window in the process of mining potential customers. Figure 5 It includes the "observation period" on the left and the "performance period" on the right. The observation period can be defined as the past 12 months, which is the time window used by the model to analyze and learn the customer's trading behavior, investment pattern and personal characteristics. During this period, the model will be trained based on the customer's historical data (such as purchasing behavior, account activity, personal attributes, etc.) to understand and predict the customer's behavior patterns. The performance period can be set within the next month after the end of the observation period. This is the time window for the model to predict whether the customer will purchase financial products. During the performance period, the model will predict whether the customer will purchase funds in the next month based on the patterns and rules learned during the observation period, thereby potentially classifying the customer. The observation time point is the dividing line between the observation period and the performance period, and is usually taken as the end of a month.

[0113] In actual production and life, the customer base managed by non-account managers of financial institutions is very large. For example, a certain bank has a total of about 700 million non-account customers (i.e., the customer base managed by non-account managers mentioned above). Only 190,000 customers who purchased non-fixed investment funds within the observation period of one month were deduplicated. Compared with the customer base that did not purchase, the data is extremely unbalanced. After data analysis, it was found that 99% of the customers in the customer base were between 18 and 65 years old, and 95% of the customers logged into mobile banking more than or equal to once in the past year. Therefore, all samples can be filtered according to the above data preprocessing (e.g., the above data deletion operation), data screening (e.g., based on information value, random forest algorithm and other screening methods), etc., which will not reduce the number of positive samples, but will effectively reduce the number of negative samples, so as to alleviate the problem of extremely unbalanced data in the sample data set (e.g., the above first sample set and the second sample set).

[0114] Optionally, in this embodiment 1, Figure 6 It is a flowchart of ensemble learning based on primary learners and secondary learners. In the primary learner stage, model training is divided into three independent but related parts, each of which is managed by a different model (Model 1, Model 2, Model 3). Model 1, Model 2, and Model 3 can be different machine learning models such as LightGBM, random forest, or decision tree. Each model will be trained with the same training set to obtain its own training results. After training, each model will be predicted on the test set to obtain its own test results. These test results are usually evaluated, such as calculating accuracy, recall, precision, etc., to measure the performance of the primary learner.

[0115] In the secondary learner stage, its task is to perform higher-level learning and prediction based on the output of the primary learner and the features of the original training set. Unlike the traditional ensemble learning model (Stacking model), the original training set is also used as part of the input of the meta-learner, which helps the meta-learner not only learn the output of the primary learner, but also directly use the features of the original data, so as to better understand the relationship between the model prediction and the actual behavior of customers.

[0116] The training results of Model 1, Model 2, and Model 3 are combined with the first training set and label information to form a new training set, namely the second training set mentioned above. The test results of Model 1, Model 2, and Model 3 are combined to form a new test set, namely the second test set mentioned above. The second test set is used to evaluate the performance of the secondary learner on unseen data after training.

[0117] Finally, the meta-learner (usually a logistic regression model) will be trained using the second training set to learn how to combine the predictions of the primary learner and the original features to make more accurate predictions of fund purchases. After training, the meta-learner will be evaluated using the second test set to ensure its generalization ability.

[0118] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0119] In summary, the method for predicting potential customers provided in the embodiment of the present application divides customers according to transaction information within a first time period to obtain a first category of customers and a second category of customers, wherein the first category of customers are customers who have not traded the target financial product within the first time period, and the second category of customers are customers who have traded the target financial product within the first time period; obtains first transaction information of the first category of customers and second transaction information of the second category of customers from the transaction information; performs feature extraction on the first transaction information to obtain first feature information, and performs feature extraction on the second transaction information to obtain second feature information; inputs the first feature information into a first prediction model to predict potential customers in the first category of customers, inputs the second feature information into a second prediction model to predict potential customers in the second category of customers, and pushes product information of the target financial product to potential customers in the first category of customers and potential customers in the second category of customers, thereby solving the problem in the related art that when a machine learning model is used to identify potential customers who purchase financial products, the sample information corresponding to transaction information of different categories of customers is unbalanced, resulting in inaccurate model output results, thereby affecting the accurate push of product information.

[0120] By extracting features from customers' transaction data and using the extracted feature information to train prediction models, it is possible to predict the likelihood of each customer becoming a potential customer of the target financial product, allowing banks to accurately and effectively identify customers with high purchasing potential. At the same time, by classifying non-client manager accounts into new and old customers based on transaction behavior, financial institutions can narrow the scope of recommended customers, reduce the cost and waste of resources of blindly reaching a large number of non-target customers, and push different product information to different types of customers, thereby improving the accuracy of financial product information push and customer satisfaction.

[0121] Example 2

[0122] The embodiment of the present application also provides a potential customer prediction device. It should be noted that the potential customer prediction device of the embodiment of the present application can be used to execute the potential customer prediction method provided by the embodiment of the present application. The potential customer prediction device provided by the embodiment of the present application is introduced below.

[0123] According to an embodiment of the present application, a device for implementing the above-mentioned potential customer prediction method is also provided. Figure 7 As shown, the device comprises:

[0124] Specifically, the classification unit 701 is used to classify customers according to transaction information within the first time period to obtain first-category customers and second-category customers, wherein the first-category customers are customers who have not traded the target financial product within the first time period, and the second-category customers are customers who have traded the target financial product within the first time period.

[0125] The acquisition unit 702 is used to acquire first transaction information of the first type of customers and second transaction information of the second type of customers from the transaction information.

[0126] The first extraction unit 703 is used to perform feature extraction on the first transaction information to obtain first feature information, and perform feature extraction on the second transaction information to obtain second feature information.

[0127] The push unit 704 is used to input the first characteristic information into the first prediction model to predict potential customers among the first category of customers, input the second characteristic information into the second prediction model to predict potential customers among the second category of customers, and push product information of the target financial product to the potential customers among the first category of customers and the potential customers among the second category of customers.

[0128] The prediction device for potential customers provided in the embodiment of the present application divides customers according to transaction information in a first time period through a dividing unit 701 to obtain a first category of customers and a second category of customers, wherein the first category of customers are customers who have not traded the target financial product in the first time period, and the second category of customers are customers who have traded the target financial product in the first time period; the acquisition unit 702 acquires first transaction information of the first category of customers and second transaction information of the second category of customers from the transaction information; the first extraction unit 703 performs feature extraction on the first transaction information to obtain first feature information, and performs feature extraction on the second transaction information to obtain second feature information; the push unit 704 inputs the first feature information into the first prediction model to predict potential customers in the first category of customers, inputs the second feature information into the second prediction model to predict potential customers in the second category of customers, and pushes product information of the target financial product to the potential customers in the first category of customers and the potential customers in the second category of customers, thereby solving the problem in the related art that when a machine learning model is used to identify potential customers who purchase financial products, the sample information corresponding to different categories of customer transaction information is unbalanced, resulting in inaccurate model output results, thereby affecting the accurate push of product information.

[0129] By extracting features from customers' transaction data and using the extracted feature information to train prediction models, it is possible to predict the likelihood of each customer becoming a potential customer of the target financial product, allowing banks to accurately and effectively identify customers with high purchasing potential. At the same time, by classifying non-client manager accounts into new and old customers based on transaction behavior, financial institutions can narrow the scope of recommended customers, reduce the cost and waste of resources of blindly reaching a large number of non-target customers, and push different product information to different types of customers, thereby improving the accuracy of financial product information push and customer satisfaction.

[0130] Optionally, in the potential customer prediction device provided in Example 2 of the present application, the above-mentioned first extraction unit 703 includes: a statistical subunit, which is used to count the first type of variables in the first transaction information according to different time dimensions to obtain the first type of data, wherein the first type of variables represent variables represented by a single numerical value; a first calculation subunit, which is used to calculate the second type of variables in the first transaction information using an unsupervised binning algorithm to obtain the second type of data, wherein the second type of variables represent variables represented by numerical values ​​within an interval range; a second calculation subunit, which is used to calculate multiple variables in the first transaction information to obtain the third type of data, wherein the multiple variables include: first type of variables, second type of variables; a first determination subunit, which is used to determine the first feature information corresponding to the first transaction information based on the first type of data, the second type of data and the third type of data.

[0131] Optionally, in the potential customer prediction device provided in Example 2 of the present application, the above-mentioned device also includes: a collection unit, which is used to collect historical transaction information of historical customers in the financial institution before inputting the first feature information into the first prediction model to predict potential customers among the first category of customers and inputting the second feature information into the second prediction model to predict potential customers among the second category of customers; a construction unit, which is used to construct a first sample set and a second sample set based on the historical transaction information; wherein the time when the historical customers in the first sample set and the second sample set purchase financial products are different; a second extraction unit, which is used to perform feature extraction on the first sample set to obtain third feature information, and to perform feature extraction on the second sample set to obtain fourth feature information; a processing unit, which is used to perform data preprocessing on the third feature information to obtain processed third feature information, and to perform data preprocessing on the fourth feature information to obtain processed fourth feature information; a training unit, which is used to train the preset model using the processed third feature information to obtain the first prediction model, and to train the preset model using the processed fourth feature information to obtain the second prediction model.

[0132] Optionally, in the prediction device for potential customers provided in Example 2 of the present application, the above-mentioned construction unit includes: a second determination subunit, used to determine a first positive sample based on historical transaction information belonging to a third category of customers, determine a first negative sample based on historical transaction information belonging to a fourth category of customers, and construct a first sample set based on the first positive sample and the first negative sample, wherein the third category of customers are customers who have not traded the target financial product in the second time period and have traded the target financial product in the third time period, the fourth category of customers are customers who have not traded the target financial product in the second time period and have not traded the target financial product in the third time period, the third time period is later than the second time period, and the first positive sample is a first negative sample. The ratio between the number of the first positive samples and the number of the first negative samples is a first preset ratio; the third determination subunit is used to determine the second positive sample based on the historical transaction information belonging to the fifth category of customers, determine the second negative sample based on the historical transaction information belonging to the sixth category of customers, and construct a second sample set based on the second positive sample and the second negative sample, wherein the fifth category of customers are customers who have traded the target financial product within the second time period and have traded the target financial product within the third time period, the sixth category of customers are customers who have traded the target financial product within the second time period and have traded the target financial product within the third time period, and the ratio between the number of the second positive samples and the number of the second negative samples is a second preset ratio.

[0133] Optionally, in the potential customer prediction device provided in Example 2 of the present application, the above-mentioned training unit includes: a division subunit, used to divide the processed third feature information into a first training set and a first test set; a first training subunit, used to respectively use the first training set to train N first preset models, to obtain N third models, and the output results of the N first preset models; a first construction subunit, used to respectively use the first test set to test the N third models, to obtain the output results of the N third models, and to construct a second test set based on the output results of the N third models; a second construction subunit, used to construct a second training set based on the output results of the N first preset models and the first training set; a second training subunit, used to train the second preset model using the second training set, and to optimize the trained second preset model using the second test set to obtain a first prediction model.

[0134] Optionally, in the potential customer prediction device provided in Example 2 of the present application, the above-mentioned processing unit includes: a filling subunit, used to determine the missing information in the third characteristic information, and fill the missing information in the third characteristic information according to the data type of the missing information to obtain the fifth characteristic information, wherein the data type of the missing information includes at least one of the following: transaction behavior type, time type, numerical type; a deleting subunit, used to delete the preset character string in the fifth characteristic information to obtain the sixth characteristic information, wherein the preset character string includes at least: a character string corresponding to repeated information, and a character string corresponding to non-semantic information; a screening subunit, used to screen key features in the sixth characteristic information based on the target indicator value to obtain the processed third characteristic information, wherein the calculation method of the target indicator value includes at least: evidence weight, information value, random forest algorithm.

[0135] Optionally, in the potential customer prediction device provided in Example 2 of the present application, the first transaction information of the first type of customer mentioned above includes at least the following information: basic information of the customer, asset information, liability information, transaction behavior information, and client behavior data.

[0136] It should be noted that the above-mentioned division unit 701, acquisition unit 702, first extraction unit 703 and push unit 704 correspond to steps S201 to S204 in Example 1, and the two modules and the corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in the above-mentioned Example 1. It should be noted that the above-mentioned modules or units can be hardware components or software components stored in a memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102 potential customer prediction, ..., 102n), and the above-mentioned modules can also be run in the computer terminal 10 provided in Example 1 as part of the device.

[0137] Example 3

[0138] An embodiment of the present application may provide an electronic device, Figure 8 is a structural block diagram of an electronic device according to an embodiment of the present application. Figure 8 As shown, the electronic device may include: one or more ( Figure 8 (only one is shown) processor 802, memory 804, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0139] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0140] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: divide the customers according to the transaction information in the first time period to obtain a first category of customers and a second category of customers, wherein the first category of customers are customers who have not traded the target financial product in the first time period, and the second category of customers are customers who have traded the target financial product in the first time period; obtain the first transaction information of the first category of customers and the second transaction information of the second category of customers from the transaction information; perform feature extraction on the first transaction information to obtain the first feature information, and perform feature extraction on the second transaction information to obtain the second feature information; input the first feature information into the first prediction model to predict potential customers among the first category of customers, input the second feature information into the second prediction model to predict potential customers among the second category of customers, and push product information of the target financial product to potential customers in the first category of customers and potential customers in the second category of customers.

[0141] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: extract features from the first transaction information to obtain first feature information, including: performing statistics on the first type of variables in the first transaction information according to different time dimensions to obtain first type of data, wherein the first type of variables represent variables represented by a single numerical value; using an unsupervised binning algorithm to calculate the second type of variables in the first transaction information to obtain second type of data, wherein the second type of variables represent variables represented by numerical values ​​within an interval; calculating multiple variables in the first transaction information to obtain third type of data, wherein the multiple variables include: first type of variables, second type of variables; determining the first feature information corresponding to the first transaction information based on the first type of data, the second type of data, and the third type of data.

[0142] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: before inputting the first feature information into the first prediction model to predict potential customers among the first category of customers, and inputting the second feature information into the second prediction model to predict potential customers among the second category of customers, the above method also includes: collecting historical transaction information of historical customers in the financial institution; constructing a first sample set and a second sample set based on the historical transaction information; wherein the time when the historical customers in the first sample set and the second sample set purchase financial products is different; performing feature extraction on the first sample set to obtain third feature information, and performing feature extraction on the second sample set to obtain fourth feature information; performing data preprocessing on the third feature information to obtain processed third feature information, and performing data preprocessing on the fourth feature information to obtain processed fourth feature information; using the processed third feature information to train the preset model to obtain the first prediction model, and using the processed fourth feature information to train the preset model to obtain the second prediction model.

[0143] The processor can call the information and application program stored in the memory through the transmission device to perform the following steps: construct a first sample set and a second sample set based on the historical transaction information, including: determining a first positive sample based on the historical transaction information belonging to the third category of customers, determining a first negative sample based on the historical transaction information belonging to the fourth category of customers, and constructing the first sample set based on the first positive sample and the first negative sample, wherein the third category of customers are customers who have not traded the target financial product in the second time period and have traded the target financial product in the third time period, the fourth category of customers are customers who have not traded the target financial product in the second time period and have not traded the target financial product in the third time period, and the third time period Later than the second time period, the ratio between the number of first positive samples and the number of first negative samples is a first preset ratio; the second positive sample is determined based on the historical transaction information of the fifth category of customers, the second negative sample is determined based on the historical transaction information of the sixth category of customers, and the second sample set is constructed based on the second positive sample and the second negative sample, wherein the fifth category of customers are customers who have traded the target financial product in the second time period and the target financial product in the third time period, the sixth category of customers are customers who have traded the target financial product in the second time period and the target financial product in the third time period, and the ratio between the number of second positive samples and the number of second negative samples is a second preset ratio.

[0144] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: use the processed third feature information to train the preset model to obtain a first prediction model, including: dividing the processed third feature information into a first training set and a first test set; use the first training set to train N first preset models respectively to obtain N third models, and the output results of the N first preset models; use the first test set to test the N third models respectively to obtain the output results of the N third models, and construct a second test set based on the output results of the N third models; construct a second training set based on the output results of the N first preset models and the first training set; use the second training set to train the second preset model, and use the second test set to optimize the trained second preset model to obtain the first prediction model.

[0145] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: perform data preprocessing on the third characteristic information to obtain processed third characteristic information, including: determining the missing information in the third characteristic information, filling the missing information in the third characteristic information according to the data type of the missing information, to obtain the fifth characteristic information, wherein the data type of the missing information includes at least one of the following: transaction behavior type, time type, numerical type; deleting the preset character string in the fifth characteristic information to obtain the sixth characteristic information, wherein the preset character string includes at least: a character string corresponding to repeated information, a character string corresponding to non-semantic information; based on the target indicator value, screening key features in the sixth characteristic information to obtain the processed third characteristic information, wherein the calculation method of the target indicator value includes at least: evidence weight, information value, random forest algorithm.

[0146] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: The first transaction information of the first type of customer includes at least the following information: the customer's basic information, asset information, liability information, transaction behavior information, and client behavior data.

[0147] It can be understood by those skilled in the art that Figure 8 The structure shown is for illustration only, and the electronic device may also be a terminal device such as a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, etc. Figure 8 The structure of the electronic device is not limited. Figure 8 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 8 Different configurations shown.

[0148] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0149] Example 4

[0150] The embodiment of the present application further provides a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the potential customer prediction method provided in the first embodiment.

[0151] Optionally, in this embodiment, the above-mentioned storage medium can be located in any one of the computer terminals in the computer terminal group in the computer network, or in any one of the mobile terminals in the mobile terminal group. The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing the program of the prediction method steps of potential customers. The above-mentioned serial numbers of the embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments. In the above-mentioned embodiments of the present application, the description of each embodiment has its own emphasis. For the parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0152] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0153] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0154] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0155] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0156] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for predicting potential customers, characterized in that: include: Dividing customers according to transaction information in a first time period to obtain a first category of customers and a second category of customers, wherein the first category of customers are customers who have not traded a target financial product in the first time period, and the second category of customers are customers who have traded a target financial product in the first time period; Acquire first transaction information of the first category of customers and second transaction information of the second category of customers from the transaction information; Extracting features from the first transaction information to obtain first feature information, and extracting features from the second transaction information to obtain second feature information; The first feature information is input into a first prediction model to predict potential customers among the first category of customers, the second feature information is input into a second prediction model to predict potential customers among the second category of customers, and product information of the target financial product is pushed to potential customers among the first category of customers and potential customers among the second category of customers.

2. The method according to claim 1, characterized in that Extracting features from the first transaction information to obtain first feature information includes: Performing statistics on first-category variables in the first transaction information according to different time dimensions to obtain first-category data, wherein the first-category variables represent variables represented by a single value; Using an unsupervised binning algorithm to calculate the second type of variables in the first transaction information to obtain second type of data, wherein the second type of variables represent variables represented by values ​​within an interval range; Calculating a plurality of variables in the first transaction information to obtain third-category data, wherein the plurality of variables include: the first-category variables and the second-category variables; The first characteristic information corresponding to the first transaction information is determined based on the first category of data, the second category of data and the third category of data.

3. The method according to claim 1, characterized in that Before inputting the first feature information into a first prediction model to predict potential customers among the first category of customers and inputting the second feature information into a second prediction model to predict potential customers among the second category of customers, the method further includes: Collect historical transaction information of historical customers in financial institutions; Constructing a first sample set and a second sample set based on the historical transaction information; wherein the time when the historical customers purchased financial products in the first sample set and the second sample set are different; Extracting features from the first sample set to obtain third feature information, and extracting features from the second sample set to obtain fourth feature information; Performing data preprocessing on the third characteristic information to obtain processed third characteristic information, and performing data preprocessing on the fourth characteristic information to obtain processed fourth characteristic information; The processed third feature information is used to train the preset model to obtain the first prediction model, and the processed fourth feature information is used to train the preset model to obtain the second prediction model.

4. The method according to claim 3, characterized in that Constructing a first sample set and a second sample set according to the historical transaction information includes: Determine a first positive sample based on the historical transaction information of the third category of customers, determine a first negative sample based on the historical transaction information of the fourth category of customers, and construct the first sample set based on the first positive sample and the first negative sample, wherein the third category of customers are customers who have not traded the target financial product in the second time period and have traded the target financial product in the third time period, and the fourth category of customers are customers who have not traded the target financial product in the second time period and have not traded the target financial product in the third time period, the third time period is later than the second time period, and the ratio between the number of the first positive samples and the number of the first negative samples is a first preset ratio; A second positive sample is determined based on the historical transaction information of the fifth category of customers, a second negative sample is determined based on the historical transaction information of the sixth category of customers, and the second sample set is constructed based on the second positive sample and the second negative sample, wherein the fifth category of customers are customers who have traded the target financial product within the second time period and the target financial product within the third time period, and the sixth category of customers are customers who have traded the target financial product within the second time period and the target financial product within the third time period, and the ratio between the number of the second positive samples and the number of the second negative samples is a second preset ratio.

5. The method according to claim 3, characterized in that: The processed third feature information is used to train a preset model to obtain the first prediction model, including: Dividing the processed third feature information into a first training set and a first test set; Using the first training set to train N first preset models respectively, to obtain N third models and output results of the N first preset models; Using the first test set to test the N third models respectively, to obtain output results of the N third models, and to construct a second test set according to the output results of the N third models; Constructing a second training set according to the output results of the N first preset models and the first training set; The second preset model is trained using the second training set, and the trained second preset model is optimized using the second test set to obtain the first prediction model.

6. The method according to claim 3, characterized in that Performing data preprocessing on the third characteristic information to obtain processed third characteristic information includes: Determine the missing information in the third characteristic information, and fill the missing information in the third characteristic information according to the data type of the missing information to obtain fifth characteristic information, wherein the data type of the missing information includes at least one of the following: transaction behavior type, time type, and value type; Deleting the preset character string in the fifth characteristic information to obtain the sixth characteristic information, wherein the preset character string at least includes: a character string corresponding to repeated information and a character string corresponding to non-semantic information; Based on the target indicator value, key features are screened in the sixth feature information to obtain the processed third feature information, wherein the calculation method of the target indicator value at least includes: evidence weight, information value, and random forest algorithm.

7. The method according to claim 1, characterized in that The first transaction information of the first type of customers includes at least the following information: basic information of the customers, asset information, liability information, transaction behavior information, and client behavior data.

8. A potential customer prediction device, characterized in that: include: a classification unit, configured to classify customers according to transaction information in a first time period to obtain a first category of customers and a second category of customers, wherein the first category of customers are customers who have not traded a target financial product in the first time period, and the second category of customers are customers who have traded a target financial product in the first time period; an acquiring unit, configured to acquire first transaction information of the first category of customers and second transaction information of the second category of customers from the transaction information; a first extraction unit, configured to extract features from the first transaction information to obtain first feature information, and to extract features from the second transaction information to obtain second feature information; A push unit is used to input the first feature information into a first prediction model to predict potential customers among the first category of customers, input the second feature information into a second prediction model to predict potential customers among the second category of customers, and push product information of the target financial product to potential customers among the first category of customers and potential customers among the second category of customers.

9. An electronic device, characterized in that: include: A memory storing an executable program; A processor, configured to run the program, wherein the program executes the method according to any one of claims 1 to 7 when running.

10. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images