Information processing method and device, electronic equipment, storage medium and product

Abstract

The invention provides an information processing method which can be applied to the technical field of big data. The information processing method comprises the steps that target information of a target customer is acquired, the target information comprises multiple pieces of combined feature information, and each piece of combined feature information is obtained by combining at least two pieces of information in multiple pieces of personal information and multiple pieces of transaction information of the target customer; a pre-trained information processing model is utilized, according to the target information, at least one probability value, corresponding to each category in the at least one category, of the target client is determined, and the pre-trained information processing model is obtained by training an initial information processing model based on the training data set; and determining a category corresponding to the target customer from the at least one category according to the at least one probability value. The invention further provides an information processing method and device, electronic equipment, a storage medium and a product.

Description

Technical Field

[0001] This application relates to the field of big data technology, specifically to an information processing method, apparatus, electronic device, storage medium, and product. Background Technology

[0002] To provide targeted and personalized services to customers, banks need to classify their customers. Once customer categories are determined, tailored services can be offered to different categories. However, current technologies rely on manual rules to determine customer categories, making it impossible to accurately identify the category based solely on customer information. Summary of the Invention

[0003] In view of the above problems, this application provides an information processing method, apparatus, electronic device, storage medium and product.

[0004] According to a first aspect of this application, an information processing method is provided, the method comprising: acquiring target information of a target customer, wherein the target information includes multiple combined feature information, each combined feature information being obtained by combining at least two types of information from multiple personal information and multiple transaction information of the target customer; using a pre-trained information processing model, determining at least one probability value corresponding to each of at least one category of the target customer based on the target information, wherein the pre-trained information processing model is trained by an initial information processing model based on a training dataset; and determining the category corresponding to the target customer from at least one category based on the at least one probability value.

[0005] According to an embodiment of this application, the method further includes: calculating the negative gradient of the loss function of the initial information processing model with respect to the model prediction based on the training dataset, and obtaining a gradient vector; constructing a feature histogram based on the gradient vector after feature binning for each feature in the training dataset; calculating the split gain of multiple candidate split points based on the feature histogram, and determining the candidate split point with the largest split gain as the target split point; updating the optimal value of each leaf node based on the target split point and the gradient vector; optimizing the current tree structure using a preset learning rate, and pruning the optimized tree structure to obtain a pruned tree structure; iteratively executing the above steps until a preset number of iterations is reached or the loss function converges, thereby obtaining a pre-trained information processing model.

[0006] According to embodiments of this application, the training dataset includes multiple training combination feature data, each of which is obtained by combining at least two types of data from multiple training personal data and multiple training transaction data.

[0007] According to an embodiment of this application, the method further includes: acquiring multiple training personal data and multiple training transaction data, wherein each type of data includes multiple training data; converting non-numerical data in the multiple training personal data and multiple training transaction data into scalar feature data to obtain converted multiple training personal data and converted multiple training transaction data; and combining at least two types of data from the converted multiple training personal data and converted multiple training transaction data to obtain a training dataset.

[0008] According to embodiments of this application, the data types of training combined feature data include scalar values ​​and vectors.

[0009] According to embodiments of this application, the feature parameters of the loss function include various training combination feature information, target personal information, and target transaction information.

[0010] According to embodiments of this application, the method further includes: acquiring multiple initial personal data and multiple initial transaction data; performing missing value processing, deduplication processing, and outlier processing on the multiple initial personal data and multiple initial transaction data to obtain multiple training personal data and multiple training transaction data.

[0011] A second aspect of this application provides an information processing apparatus, comprising: an acquisition module for acquiring target information of a target customer, wherein the target information includes multiple combined feature information, each combined feature information being obtained by combining at least two types of information from multiple personal information and multiple transaction information of the target customer; a first determination module for determining, based on the target information, at least one probability value corresponding to each of at least one category of the target customer using a pre-trained information processing model, wherein the pre-trained information processing model is trained by an initial information processing model based on a training dataset; and a second determination module for determining the category corresponding to the target customer from at least one category based on at least one probability value.

[0012] A third aspect of this application provides an electronic device comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method described above.

[0013] A fourth aspect of this application also provides a computer-readable storage medium having a computer program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0014] The fifth aspect of this application also provides a computer program product, including a computer program or instructions that, when executed by a processor, implement the steps of the above-described method. Attached Figure Description

[0015] The above-mentioned contents, other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0016] Figure 1 The illustrations depict application scenarios of information processing methods, apparatuses, devices, media, and program products according to embodiments of this application.

[0017] Figure 2 A flowchart illustrating an information processing method according to an embodiment of this application is shown schematically.

[0018] Figure 3 A flowchart illustrating an information processing method according to another embodiment of this application is shown schematically;

[0019] Figure 4 A schematic diagram of a training dataset according to an embodiment of this application is shown.

[0020] Figure 5 This schematically illustrates a structural block diagram of an information processing apparatus according to an embodiment of the present application; and

[0021] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing an information processing method according to an embodiment of this application. Detailed Implementation

[0022] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0023] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0024] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0025] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0026] In the technical solution of this application, the user information (including but not limited to user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, and displayed data) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with relevant laws, regulations, and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse.

[0027] In scenarios involving automated decision-making using personal information, the methods, devices, and systems provided in this application all offer users corresponding entry points for choosing to agree to or reject the automated decision-making results. If the user chooses to reject, the process proceeds to the expert decision-making stage. Here, "automated decision-making" refers to the activity of automatically analyzing and evaluating an individual's behavioral habits, interests, or economic, health, and credit status through computer programs, and then making a decision. Here, "expert decision-making" refers to the activity of making decisions by personnel who specialize in a particular field, possess specialized experience, knowledge, and skills, and have reached a certain level of professional expertise.

[0028] The embodiments of this application provide an information processing method.

[0029] Figure 1 The illustrations depict application scenarios of information processing methods, apparatuses, devices, media, and program products according to embodiments of this application.

[0030] like Figure 1 As shown, application scenario 100 according to this embodiment may include the field of big data technology. Network 104 is used as a medium to provide a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. Network 104 may include various connection types, such as wired or wireless communication links or fiber optic cables, etc.

[0031] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0032] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0033] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0034] It should be noted that the information processing method provided in this application embodiment can generally be executed by server 105. Correspondingly, the information processing device provided in this application embodiment can generally be located in server 105. The information processing method provided in this application embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the information processing device provided in this application embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105.

[0035] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0036] The following will be based on Figure 1 The described scene, through Figures 2-6 The information processing method according to the embodiments of this application will be described in detail.

[0037] Figure 2 A flowchart illustrating an information processing method according to an embodiment of this application is shown.

[0038] like Figure 2 As shown, the information processing method of this embodiment includes operations S210 to S230, and the information processing method can be executed by a server.

[0039] In operation S210, target information of the target customer is obtained. The target information includes multiple combined feature information. Each combined feature information is obtained by combining at least two types of information from multiple personal information and multiple transaction information of the target customer.

[0040] In the embodiments of this application, it is necessary to determine the category of the target customer based on the target customer's target information. A target customer refers to a customer whose category needs to be defined. Various personal information of the target customer may include their personal information, financial information, credit score information, product usage information, behavioral data, geographic information, risk indicators, etc. It should be noted that the acquisition of the aforementioned personal information, financial information, credit score information, product usage information, behavioral data, geographic information, risk indicators, etc., of the target customer should comprehensively consider factors such as business objectives, regulatory compliance, and customer privacy, and the collected information must comply with relevant data privacy laws and regulations.

[0041] The target customer's transaction information may include their transaction history. For example, transaction history may include the target customer's historical transaction records, such as deposits, withdrawals, and transfers, and may also include the target customer's credit card usage, such as repayment history. It should be noted that the acquisition of the above information should comprehensively consider factors such as business objectives, regulatory compliance, and customer privacy, and the collected information must comply with relevant data privacy laws and regulations.

[0042] In embodiments of this application, the target information includes multiple combined feature information, each of which is obtained by combining at least two types of information from multiple personal information and multiple transaction information of the target customer. Each combined feature information may be obtained by combining two types of personal information of the target customer, two types of transaction information of the target customer, or one type of personal information and one type of transaction information of the target customer. Each combined feature information may also be obtained by combining three types of information, which may include one type of personal information and two types of transaction information.

[0043] In the embodiments of this application, the combination of information can be obtained by performing arithmetic operations on the scalar corresponding to each type of information.

[0044] For example, combined characteristic information can be obtained by multiplying transaction amount (average transaction amount over a specific time period) and transaction frequency (number of transactions over a specific time period). This combined characteristic information can reflect the overall transaction volume of a target customer while taking into account the frequency of transactions. A high value of this combined characteristic information can indicate an active customer with large transactions.

[0045] For example, combined feature information can be obtained by multiplying a credit score (a higher credit score indicates better creditworthiness) by the number of loans taken out. By considering the product of the credit score and loan history, customers with good credit and extensive loan experience can be identified.

[0046] For example, combined characteristic information can be obtained by dividing the transaction amount (average transaction amount over a specific period) by the account balance. This combined characteristic information can characterize the ratio of a customer's average transactions to their account balance, helping to understand the customer's capital utilization efficiency. A high value of this combined characteristic information can indicate a customer with high capital utilization efficiency.

[0047] In operation S220, a pre-trained information processing model is used to determine at least one probability value for the target customer and each of at least one category based on the target information. The pre-trained information processing model is obtained by training the initial information processing model based on the training dataset.

[0048] In embodiments of this application, target information can be input into a pre-trained information processing model to obtain probability values ​​output by the pre-trained information processing model. For example, the probability value can be expressed as... Where Y represents the category label for customer classification, X represents the target information of the target customer, and i represents different categories.

[0049] In operation S230, the category corresponding to the target customer is determined from at least one category based on at least one probability value.

[0050] In embodiments of this application, the category with the highest probability value can be selected as the category corresponding to the target customer. Classifying the target customer can involve determining a single category or multiple categories. For determining a single category, a threshold T can be set; if... If the target customer is classified into category i, then the target customer is classified into category i; otherwise, they belong to other categories. For classification problems with multiple categories, the formula can be used... The category with the highest probability value is selected as the category corresponding to the target customer.

[0051] Through embodiments of this application, target information including multiple combined feature information is input into a pre-trained information processing model to obtain probability values ​​output by the pre-trained information processing model. The category corresponding to the target customer is then determined based on these probability values. By defining multi-dimensional combined feature information, the feature representation capability of the target information can be improved, thereby increasing the accuracy of determining the category corresponding to the target customer.

[0052] Figure 3 A flowchart illustrating an information processing method according to another embodiment of this application is shown.

[0053] like Figure 3 As shown, in some embodiments, the information processing method includes operations S310 to S360, which can be executed by a server.

[0054] In operation S310, the negative gradient of the loss function of the initial information processing model with respect to the model prediction is calculated based on the training dataset, and the gradient vector is obtained.

[0055] In embodiments of this application, the information processing model can be a gradient boosting framework. The gradient boosting framework can iteratively optimize the objective loss function through gradient descent (in the form of pseudo-residuals).

[0056] In the embodiments of this application, the training dataset includes multiple sample (training client) data, comprising a preprocessed and feature-combined feature matrix X and a label y. For each sample i, the negative gradient of the loss function relative to the model prediction can be calculated using formula (1):

[0057] (1)

[0058] in, For loss function, These are the model predictions. It should be noted that initially, for regression problems, the model predictions are initialized to constants, such as the mean of the sample labels; for classification problems, they are initialized to prior probabilities.

[0059] After calculating the negative gradient for each sample, the gradient vector g = (g 1, g 2,…, g n Each sample corresponds to a gradient value.

[0060] Understandably, this step can convert the original label y and the current model prediction value F into gradient signals to guide the subsequent construction of the tree structure.

[0061] After performing feature bucketing on each feature in the training dataset using the S320, a feature histogram is constructed based on the gradient vector.

[0062] In the embodiments of this application, constructing a histogram for the leaf nodes involves binning each feature in the training dataset to form a histogram. Each bin contains multiple samples. For each bin, the gradient sum of the samples is calculated based on the gradient vector to construct a feature histogram. For example, for each feature j, its value range is divided into k discrete bins. The age feature can be divided into four bins: 0-20, 20-40, 40-60, and 60+. For each bin b, the sum of the gradients of all samples within that bin is calculated.

[0063] Understandably, this step compresses the original features and gradient information into histogram form, which can significantly reduce the amount of computation and avoid traversing all sample points.

[0064] In operation S330, the splitting gain of multiple candidate segmentation points is calculated based on the feature histogram, and the candidate segmentation point with the largest splitting gain is determined as the target segmentation point.

[0065] In the embodiments of this application, for each feature j, all possible bucket boundaries are traversed as candidate split points. For each candidate split point s, the data is divided into left and right nodes, and its split gain is calculated using formula (2):

[0066] (2)

[0067] Among them, sg l Let sg be the gradient sum of the left node. r Let sh be the gradient sum of the right node. l Let sh be the sum of the Hessian matrices of the left node. r The sum of the Hessian matrices of the right node. This is a regularization term.

[0068] Compare the gains of all candidate split points, and determine the candidate split point with the largest split gain as the target split point.

[0069] Understandably, in this step, histogram statistics are used to quickly calculate the split gain, avoiding traversal of the original data.

[0070] In operation S340, the optimal value of each leaf node is updated based on the target split point and gradient vector.

[0071] In the embodiments of this application, the optimal value of each leaf node is calculated by minimizing the objective function using formula (3):

[0072] (3)

[0073] Where sg is the sum of gradients of the target segmentation points, and sh is the sum of the Hessian matrices of the target segmentation points.

[0074] When operating S350, the current tree structure is optimized using a preset learning rate, and the optimized tree structure is pruned to obtain the pruned tree structure.

[0075] In the embodiments of this application, the complete structure of the current tree includes all split points and leaf values. Multiplying the output value of each leaf node by a preset learning rate (e.g., 0.1) can prevent a single tree from having too much influence. The optimized tree structure is pruned, each split point is checked, and its split gain is calculated and compared with a preset threshold. If the split gain after splitting is less than the preset threshold, the split is canceled (pruning), resulting in a tree structure after shrinking and pruning.

[0076] In operation S360, the above steps are executed iteratively until the preset number of iterations is reached or the loss function converges, thus obtaining the pre-trained information processing model.

[0077] In the embodiments of this application, S310-S350 are repeatedly executed until a preset number of iterations is reached or the loss function converges, thereby obtaining a pre-trained information processing model.

[0078] Through the embodiments of this application, the initial information processing model is trained based on the training dataset to obtain a pre-trained information processing model. Through this series of steps, the pre-trained information processing model can gradually build an ensemble model of weak classifiers to achieve efficient and accurate classification of target information.

[0079] In some embodiments, the training dataset includes multiple training combination feature data, each of which is obtained by combining at least two types of data from multiple training individual data and multiple training transaction data.

[0080] Figure 4 A schematic diagram of a training dataset according to an embodiment of this application is shown.

[0081] like Figure 4 As shown, the data in the training personal data can include multiple categories: category a, category b, category c, and category d. The data in the training transaction data can include multiple categories: category e, category f, category g, and category h.

[0082] In the embodiments of this application, each training combination feature data is obtained by combining at least two types of data from multiple training personal data and multiple training transaction data.

[0083] For example, training combination feature data 1 is obtained by combining category a from training individual data and category e from training transaction data.

[0084] For example, training combination feature data 2 is obtained by combining category b and category c from the training individual data.

[0085] It should be noted that, Figure 4 For illustrative purposes only, the number of categories of training personal data and training trading data is not limited to this. Figure 4 The four types shown.

[0086] In the embodiments of this application, the training personal data may include personal information, financial information, credit score information, product usage information, behavioral data, geographic information, risk indicators, and other information. It should be noted that the acquisition of the aforementioned personal information, financial information, credit score information, product usage information, behavioral data, geographic information, risk indicators, and other information should comprehensively consider factors such as business objectives, regulatory compliance, and customer privacy, and the collected information must comply with relevant data privacy laws and regulations.

[0087] Training transaction data may include the training customer's transaction history. For example, transaction history information may include the training customer's historical transaction records, such as deposits, withdrawals, and transfers. It may also include the training customer's credit card usage, such as repayment history. It should be noted that the acquisition of the above information should comprehensively consider factors such as business objectives, regulatory compliance, and customer privacy, and the collected information must comply with relevant data privacy laws and regulations.

[0088] Through the embodiments of this application, the training combination feature data in the training dataset is composed of multi-dimensional feature combination rules, which can significantly improve the feature expression ability and enhance the processing capability of the information processing model trained based on the training dataset.

[0089] In some embodiments, multiple training personal data and multiple training transaction data are acquired, wherein each data includes multiple training data; non-numerical data in the multiple training personal data and multiple training transaction data are converted into scalar feature data to obtain converted multiple training personal data and converted multiple training transaction data; at least two types of data in the converted multiple training personal data and converted multiple training transaction data are combined to obtain a training dataset.

[0090] In the embodiments of this application, since the training personal data and training transaction data contain non-numerical data, for example, for training combination feature data of region and consumption habits, the non-numerical data can be converted into scalar feature data.

[0091] For example, "Region A" corresponds to the code [1,0,0], and "Region B" corresponds to the code [0,1,0]. When the region is "Region A" and the consumption habit is an average monthly consumption of 5000 yuan, the two data can be multiplied to get [5000,0,0].

[0092] In embodiments of this application, at least two types of data from the transformed training personal data and the transformed training transaction data are combined to obtain a training dataset. Specific combination methods may include arithmetic operations.

[0093] For example, transaction amount and transaction frequency. This training combination of features can reflect a customer's overall transaction volume while taking into account transaction frequency. High values ​​may indicate active customers with large transactions.

[0094] For example, credit score and loan history. This training combination of feature data takes into account the product of credit score and loan history, which helps identify customers with good credit and more loan experience.

[0095] For example, average transaction amount / account balance. This training combination of feature data can characterize the ratio of a customer's average transactions to their account balance, helping to understand the efficiency of a customer's capital utilization.

[0096] For example, transaction time period and transaction frequency. This training combined feature data considers the product of the time period and frequency of transactions, which helps to identify customers with high transaction frequency within a specific time period, such as during holiday promotions.

[0097] For example, age and loan amount. By multiplying age by loan amount, this training dataset can reveal the loan performance of customers in different age groups and potentially uncover their loan preferences.

[0098] For example, average transaction amount and credit score. By considering the product of average transaction amount and credit score, this training combination of feature data can reveal the spending level of customers with high credit scores in average transactions, helping to understand the transaction habits of high-credit customers.

[0099] For example, account balance and transaction frequency. This training portfolio of features considers the product of account balance and transaction frequency, which can help identify changes in transaction frequency when account balances are relatively high, potentially indicating a customer's investment or spending behavior.

[0100] For example, age and average transaction amount. This training set of feature data reflects the average transaction amount of customers in different age groups, which helps to identify the impact of age on transaction amount.

[0101] For example, credit score and transaction frequency. By considering the product of credit score and transaction frequency, this training combined feature data can observe the performance of customers with different credit levels in terms of transaction frequency, which helps to understand the relationship between credit score and transaction activity.

[0102] For example, the number of loans and account balance. This training composite feature data, by multiplying the number of loans by the account balance, can reveal the impact of loans on a customer's account balance, potentially indicating the relationship between loans and savings.

[0103] For example, transaction time periods and average transaction amounts. This training combination of feature data, by considering the product of the time period in which transactions occur and the average transaction amount, can reveal the average transaction level of customers within different time periods, helping to identify periods of high transaction activity.

[0104] For example, age, credit score, and loan amount. This training combination of feature data, through triple training and combination of feature data, can comprehensively consider the relationship between age, credit score, and loan amount, which helps to uncover more complex patterns.

[0105] For example, transaction frequency, account balance, and average transaction amount. This training combination of feature data, through triple training and combination of feature data, can comprehensively consider the relationship between transaction frequency, account balance, and average transaction amount, helping to capture more transaction behavior patterns.

[0106] For example, age, transaction frequency, and credit score. This training combination of feature data, through triple training and combination of feature data, can comprehensively consider customers' age, transaction activity, and credit score, helping to discover classification patterns of different subgroups.

[0107] For example, average transaction amount, number of loans, and credit score. This training combination of feature data, through triple training and combination of feature data, can comprehensively consider the relationship between average transaction amount, number of loans, and credit score, which helps to understand the overall financial behavior of customers.

[0108] Through the embodiments of this application, the training combined feature data in the bank customer classification scenario typically outputs scalar values. Regardless of the original feature type (continuous / discrete), the combination result needs to be converted into a numerical value that can participate in mathematical operations. All features (including combined features) are input into the model in the form of a numerical matrix. By combining features, the nonlinear relationship between the original features can be revealed, which can improve the model's representation ability.

[0109] In some embodiments, the data types of the training combined feature data include scalar values ​​and vectors.

[0110] For example, age, credit score, and number of loans. If the training customer is 45 years old, has a credit score of 750, and has taken out 3 loans, the scalar value of the training combination feature data can be calculated to be 101250.

[0111] For example, age and number of transactions can be used to construct the feature vector for training combined feature data in the form of [age, number of transactions, combined feature (age and number of transactions)]. If the training customer is 35 years old and has made 12 transactions, the feature vector for training combined feature data could be [35, 12, 420].

[0112] Through the embodiments of this application, training combined feature data is input into the model in the form of a numerical matrix. The nonlinear relationship between the original features is revealed through feature combination, thereby improving the model's representation ability.

[0113] In some embodiments, the feature parameters of the loss function include various training combination feature information, target personal information, and target transaction information.

[0114] In the embodiments of this application, the target transaction information can be transaction frequency (the number of transactions within a target time period), and the target personal information can be credit rating. By selecting transaction frequency and credit rating as business-related feature parameters, they can be introduced into the model and the objective function can be adjusted.

[0115] Update the feature vector x input to the information processing model i . ,in, The training dataset includes various training combinations of feature data, where i represents the i-th customer, and TF... i CR represents the transaction frequency of the i-th user. i This represents the credit rating of the i-th user.

[0116] Update the objective function. Original objective function: .

[0117] Updated objective function: .

[0118] Update the loss function:

[0119] ,in, and These are the weighting coefficients of the business parameters, used to balance the contribution of business features to the loss function. and These are the training parameters in the loss function.

[0120] Through the embodiments of this application, by introducing business-related feature parameters into the loss function of model training, the model can pay more attention to features directly related to the business, thereby better adapting to the specific needs of bank customer classification business.

[0121] In some embodiments, the method further includes: acquiring multiple initial personal data and multiple initial transaction data; performing missing value processing, deduplication processing, and outlier processing on the multiple initial personal data and multiple initial transaction data to obtain multiple training personal data and multiple training transaction data.

[0122] In the embodiments of this application, missing values ​​refer to fields in the dataset that are empty or missing. Methods for handling missing values ​​include deletion, imputation, and interpolation. Rows or columns with few missing values ​​can be directly deleted. Imputation methods can use fixed values, mean, median, mode, etc., to fill missing values. In this embodiment, the mean can be used for imputation. Deduplication is the process of eliminating duplicate samples in the data. For duplicate data during information processing, duplicate rows in the dataset can be directly deleted. Outliers refer to values ​​in the dataset that differ significantly from normal values ​​and may have a negative impact on the model. Outlier handling methods include deletion and imputation; data that clearly deviates from the normal range can be directly deleted.

[0123] The embodiments of this application can effectively handle missing values, remove duplicate data, and process outliers, thereby improving the accuracy of the information processing process.

[0124] Based on the above information processing method, this application also provides an information processing apparatus. The following will be combined with... Figure 5 The device is described in detail.

[0125] Figure 5 A schematic block diagram of an information processing apparatus according to an embodiment of this application is shown.

[0126] like Figure 5 As shown, the information processing device 500 of this embodiment includes an acquisition module 510, a first determination module 520, a second determination module 530, and a third determination module 440.

[0127] The acquisition module 510 is used to acquire target information of a target customer, wherein the target information includes multiple combined feature information, and each combined feature information is obtained by combining at least two types of information from multiple personal information and multiple transaction information of the target customer. In one embodiment, the acquisition module 510 can be used to perform the operation S210 described above, which will not be repeated here.

[0128] The first determining module 520 is configured to use a pre-trained information processing model to determine at least one probability value corresponding to each of the at least one category of the target customer based on the target information, wherein the pre-trained information processing model is trained by an initial information processing model based on a training dataset. In one embodiment, the first determining module 520 may be used to perform the operation S220 described above, which will not be repeated here.

[0129] The second determining module 530 is configured to determine the category corresponding to the target customer from the at least one category based on the at least one probability value. In one embodiment, the second determining module 530 may be used to perform the operation S230 described above, which will not be repeated here.

[0130] According to embodiments of this application, any plurality of modules among the acquisition module 510, the first determination module 520, and the second determination module 530 can be combined into one module, or any one of these modules can be split into multiple modules. Alternatively, at least a portion of the functionality of one or more of these modules can be combined with at least a portion of the functionality of other modules and implemented in one module. According to embodiments of this application, at least one of the acquisition module 510, the first determination module 520, and the second determination module 530 can be at least partially implemented as a hardware circuit, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or implemented by any other reasonable means of integrating or packaging the circuit, or implemented in any one of the three methods of software, hardware, and firmware, or in a suitable combination of any of these. Alternatively, at least one of the acquisition module 510, the first determination module 520, and the second determination module 530 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is run.

[0131] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing an information processing method according to an embodiment of this application.

[0132] like Figure 6 As shown, an electronic device 600 according to an embodiment of this application includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage portion 608 into a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this application.

[0133] RAM 603 stores various programs and data required for the operation of electronic device 600. Processor 601, ROM 602, and RAM 603 are interconnected via bus 604. Processor 601 executes various operations of the method flow according to embodiments of this application by executing programs in ROM 602 and / or RAM 603. It should be noted that programs may also be stored in one or more memories other than ROM 602 and RAM 603. Processor 601 may also execute various operations of the method flow according to embodiments of this application by executing programs stored in one or more memories.

[0134] According to embodiments of this application, the electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to a bus 604. The electronic device 600 may also include one or more of the following components connected to the input / output (I / O) interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 610 as needed so that computer programs read from it can be installed into the storage section 608 as needed.

[0135] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0136] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this application, the computer-readable storage medium may include ROM 602 and / or RAM 603 and / or one or more memories other than ROM 602 and RAM 603 described above.

[0137] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to enable the computer system to implement the information processing methods provided in the embodiments of this application.

[0138] When the computer program is executed by the processor 601, it performs the functions defined in the system / apparatus of this application embodiment. According to the embodiments of this application, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0139] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and downloaded and installed via the communication section 609, and / or installed from the removable medium 611. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0140] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, it performs the functions defined in the system of this application embodiment. According to the embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0141] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0142] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0143] Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

Claims

1. An information processing method, characterized in that, The method includes: Obtain target information of target customers, wherein the target information includes multiple combined feature information, and each combined feature information is obtained by combining at least two types of information from multiple personal information and multiple transaction information of the target customer; Using a pre-trained information processing model, at least one probability value is determined for each of the target customers in at least one category, based on the target information, wherein the pre-trained information processing model is trained from an initial information processing model based on a training dataset; The category corresponding to the target customer is determined from the at least one category based on the at least one probability value.

2. The method according to claim 1, characterized in that, The method further includes: The negative gradient of the loss function of the initial information processing model with respect to the model prediction is calculated based on the training dataset to obtain the gradient vector; After binning each feature in the training dataset, a feature histogram is constructed based on the gradient vector; Calculate the splitting gain of multiple candidate segmentation points based on the feature histogram, and determine the candidate segmentation point with the largest splitting gain as the target segmentation point; Update the optimal value of each leaf node based on the target segmentation point and the gradient vector; The current tree structure is optimized using a preset learning rate, and the optimized tree structure is pruned to obtain the pruned tree structure. The above steps are executed iteratively until the preset number of iterations is reached or the loss function converges, thus obtaining the pre-trained information processing model.

3. The method according to claim 1, characterized in that, The training dataset includes multiple training combination feature data, each of which is obtained by combining at least two types of data from multiple training personal data and multiple training transaction data.

4. The method according to claim 3, characterized in that, The method further includes: Acquire multiple types of training personal data and multiple types of training transaction data, where each type of data includes multiple training data; Non-numerical data from various training personal data and various training transaction data are converted into scalar feature data to obtain the converted training personal data and the converted training transaction data. The training dataset is obtained by combining at least two types of data from the transformed multiple training personal data and the transformed multiple training transaction data.

5. The method according to claim 3, characterized in that, The data types of the training combination feature data include scalar values ​​and vectors.

6. The method according to claim 3, characterized in that, The feature parameters of the loss function include the feature information of the various training combinations, the target's personal information, and the target's transaction information.

7. The method according to claim 4, characterized in that, The method further includes: Acquire various initial personal data and various initial transaction data; Missing value processing, deduplication processing, and outlier processing are performed on the various initial personal data and various initial transaction data to obtain the various training personal data and the various training transaction data.

8. An information processing device, characterized in that, The device includes: The acquisition module is used to acquire target information of target customers, wherein the target information includes multiple combined feature information, and each combined feature information is obtained by combining at least two types of information from multiple personal information and multiple transaction information of the target customer; The first determining module is used to determine, based on the target information, at least one probability value corresponding to each of the target customers and at least one category, using a pre-trained information processing model, wherein the pre-trained information processing model is trained by an initial information processing model based on a training dataset. The second determining module is used to determine the category corresponding to the target customer from the at least one category based on the at least one probability value.

9. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 7.

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

Images