Object value determination method

Abstract

The embodiment of the invention discloses an object value determination method. The method comprises the steps that first object data are acquired, the target probability that a target object handles a target service is predicted according to the first object data and a service handling prediction model, and the first object data comprise data of the target object under association features associated with the target service; determining an object value of the target object according to the first object data and the target probability; wherein the business handling prediction model is obtained by training at least one associated sample, and the associated sample comprises data of the sample object under the associated features. According to the technical scheme provided by the embodiment of the invention, the object value can be efficiently and accurately determined.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of data processing technology, and in particular to a method for determining the value of an object. Background Art

[0002] Determining the value of an object can help discover the potential value that the object may generate in the future, and facilitate the adoption of service measures corresponding to the value of the object, thereby improving the stickiness and loyalty of the object and achieving the company's long-term profit goals.

[0003] Currently, the value of an object is determined manually. However, manual determination of the value of an object is difficult to perform efficiently and accurately. Summary of the Invention

[0004] The embodiment of the present invention provides a method for determining the value of an object, so as to achieve efficient and accurate determination of the value of an object.

[0005] According to one aspect of the present invention, a method for determining an object value is provided, which may include:

[0006] Acquire first object data, and predict a target probability of a target object handling a target business based on the first object data and a business handling prediction model, wherein the first object data includes data of the target object under associated features associated with the target business;

[0007] determining an object value of the target object based on the first object data and the target probability;

[0008] The business processing prediction model is trained based on at least one associated sample, and the associated sample includes data of the sample object under associated features.

[0009] The technical solution of the embodiment of the present invention obtains first object data, and predicts the target probability of the target object handling the target business based on the first object data and the business handling prediction model, so as to facilitate the subsequent determination of the object value through the target probability, wherein the first object data includes the data of the target object under the associated features associated with the target business, and then the object value of the target object is determined based on the first object data and the target probability to achieve the determination of the object value, wherein the business handling prediction model is obtained by training based on at least one associated sample, and the associated sample includes the data of the sample object under the associated features. The above technical solution does not need to rely on manual labor, and the business prediction model obtained by training the associated samples including the data under the associated features associated with the target business can make the business prediction model more targeted and automatically predict the probability of handling the target business, thereby efficiently and accurately predicting the target probability, and then achieving efficient and accurate determination of the object value based on the target probability.

[0010] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 is a flow chart of a method for determining object value according to an embodiment of the present invention;

[0013] Figure 2 is a flow chart of another object value determination method provided according to an embodiment of the present invention;

[0014] Figure 3 This is a flow chart of obtaining a wide table template in another object value determination method provided by an embodiment of the present invention;

[0015] Figure 4 is a flowchart of an optional example of another object value determination method provided by an embodiment of the present invention;

[0016] Figure 5 This is a structural block diagram of a device for determining object value according to an embodiment of the present invention;

[0017] Figure 6 It is a structural diagram of an electronic device for implementing the object value determination method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0019] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. The situations of "target", "original", etc. are similar and will not be repeated here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products or devices.

[0020] It is understandable that before using the technical solutions disclosed in the various embodiments of this disclosure, the type, scope of use, usage scenarios, etc. of the personal information involved in this disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0021] Before introducing the embodiments of the present invention, an example description is given of the current application scenarios and implementation processes of the solutions used for determining the value of an object, as well as the reasons why it is difficult to efficiently and accurately determine the value of an object, so as to better understand why the solutions proposed in the embodiments of the present invention can achieve efficient and accurate determination of the value of an object.

[0022] After years of development, banks have entered a phase where they prioritize both existing and future business opportunities. Certain businesses are large-scale, with one currently boasting nearly 200 billion yuan in sales. In addition to exploring new sources of growth, these businesses also require exploring the potential value of their clients to enhance their ability to sustainably manage these businesses.

[0023] Object value determination usually refers to the discovery of the potential value that an object may generate in the future through the analysis and utilization of object data, as well as the facilitation of the adoption of marketing strategies and service measures corresponding to the value of the object, thereby increasing the stickiness and loyalty of the object and achieving the company's long-term profit goals.

[0024] Currently, the value of an object is manually determined by project managers and others. However, the traditional method of manually determining the value of an object faces problems such as a large amount of due diligence tasks, high professional requirements, and a long time-consuming single business approval task. Therefore, it is currently difficult to determine the value of an object efficiently and accurately.

[0025] To address this, embodiments of the present invention eliminate the need for manual effort. Instead, they utilize a business prediction model trained using associated samples of data related to the target business's associated features. This allows the business prediction model to more accurately and efficiently predict the probability of handling the target business, thereby efficiently and accurately predicting the target probability and, in turn, determining the object's value. This will be explained in detail below.

[0026] Figure 1 This is a flow chart of a method for determining object value provided in an embodiment of the present invention. This embodiment is applicable to situations involving object value determination. This method can be performed by an object value determination device provided in an embodiment of the present invention. This device can be implemented in software and / or hardware and can be integrated into an electronic device, such as various user terminals or servers.

[0027] See also Figure 1 The method of the embodiment of the present invention specifically includes the following steps:

[0028] S110. Obtain first object data, and predict the target probability of the target object handling the target business based on the first object data and the business handling prediction model, wherein the first object data includes data of the target object under associated features associated with the target business, and the business handling prediction model is trained based on at least one associated sample, and the associated sample includes data of the sample object under the associated features.

[0029] Among them, the first object data can be understood as data including the data of the target object under the associated features associated with the target business; the first object data is obtained by data mining the target object's existing data on handling the target business, or it can be provided by the target object itself.

[0030] The target object can be understood as the object whose value is determined by the demand.

[0031] It should be noted that the type, scope of use, usage scenarios, etc. of personal information involved in obtaining the first object data should be informed to the target object and authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations. The historical data and second object data in the embodiments of the present invention should all be informed to the corresponding object and authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0032] The business processing prediction model can be understood as a model used to predict the target probability of a target object processing a target business.

[0033] Associated samples can be understood as samples used to train business processing models.

[0034] The sample object can be understood as the object corresponding to the associated sample.

[0035] In an embodiment of the present invention, a business processing prediction model may be obtained by training based on at least one data association sample including a sample object under an associated feature.

[0036] In an embodiment of the present invention, the data and / or associated features of the sample object included in the associated sample under the associated features can be encoded in advance for each associated sample, the associated sample can be updated according to the encoding result, and then a business processing prediction model can be obtained by training based on at least one associated sample, so as to achieve faster training of the business processing prediction model through the encoded associated samples.

[0037] In an embodiment of the present invention, considering that extreme gradient boosting (eXtreme Gradient Boosting, XGBoost) performs well, especially when processing complex data sets with a large number of features and samples, and it has better generalization ability, it is widely used in tasks such as classification, regression and sorting, and it has high efficiency and accuracy. Therefore, the business processing prediction model can be, for example, an XGBoost model. On this basis, the XGBoost machine learning algorithm can be used to train the XGBoost model. XGBoost can improve the prediction performance of the model by constructing multiple decision trees and performing weighted combination. During the model training process, the data used to train the model can be divided into a training set and a test set. The training set can include, for example, at least one associated sample. It can also be that at least one associated sample is divided into a training set and a test set, and the training set is used to train the XGBoost model, and the test set is used to evaluate the performance of the XGBoost model. During the model training process, the parameters of the model, such as the learning rate, the depth of the tree, and the regularization parameter, can be adjusted to control the complexity of the model, reduce the risk of overfitting of the model, and thus optimize the performance of the model. The mean squared error (MSE) loss function can be used. As the loss function of the prediction regression task of the XGBoost model, where y i is the actual value output by the model, is the predicted value of the associated sample as a label, N is the number of associated samples trained using the above loss function, and L(θ) is the mean square error; in addition, the gradient boosting method can be used in the training of the XGBoost model to gradually optimize the model. Each iteration generates a new tree to fit the residual of the previous round, where is the prediction of the tth round, η is the learning rate, f t (x) is the tree model generated in round t.

[0038] The target business can be understood as a business that can be handled by the target object; the target business can be, for example, a loan business, a credit card business, or a commodity purchase business, etc.; the number of target businesses can be at least one. In this case, the first object data can be obtained for each target business, and based on the first object data and the business handling prediction model, the target probability of the target object handling the target business can be predicted, and subsequently, the object value of the target object can be determined based on the first object data and target probability corresponding to at least one target business.

[0039] The target probability can be understood as the probability that the target object handles the target business.

[0040] The associated features can be understood as features associated with the target business; the associated features can, for example, include at least one of numerical features, classification features, and time series features, etc. In the embodiment of the present invention, there is no specific limitation on the feature type of the associated features; the number of associated features can be at least one.

[0041] It can be understood that the first object data can reflect the target probability, and the target probability can reflect the object value. That is, for example, within the set time, if the target object has a high income and a high loan amount, then the target object's loan probability is high, and the target object can be considered to be a high-quality user with a relatively long life cycle and frequent renewals and a high object value. On the contrary, if the target object has a low income and a low loan amount, then the target object's loan probability is low, and the loan amount is also relatively low, then the target object can be considered to be a low-quality object with a short loan life cycle and low object value. Therefore, the first object data can be obtained, and the target probability can be predicted based on the first object data and the business processing prediction model, and then the object value can be determined subsequently based on the first object data and the target probability.

[0042] In the embodiment of the present invention, the first object data can be acquired in real time to determine the object value in real time.

[0043] S120: Determine the object value of the target object according to the first object data and the target probability.

[0044] Among them, object value can be understood as the value that can characterize the target object; object value can be expressed in a data-based or numerical form; object value can, for example, reflect at least one of the target object's current consumption capacity, business contribution level, future value potential, and business handling behavior tendencies, etc.

[0045] In the embodiment of the present invention, the object value can be determined based on the first object data and the target probability. For example, the object value can be determined based on the first object data and the target probability by the formula: Determine the object value, where CLV is the object value, M is the processing value associated with the business processing in the first object data, and the processing value can be, for example, the loan amount per period or the average amount or quantity of historical purchases, etc., and r is the target probability.

[0046] In an embodiment of the present invention, the object value can also be determined based on the first object data through at least one of regression analysis, time series analysis, decision tree and rule engine. For example, the object value can be determined through a decision tree and rule engine based on a series of "if-then" rules and the first object data. Specifically, if the first object data indicates that the target object has handled the target business of purchasing goods exceeding a certain amount in the past year, it can be determined that the object value of the target object is relatively high.

[0047] In an embodiment of the present invention, after determining the target value of the target object based on the first object data and the target probability, a value threshold may be obtained. If the target value is greater than the value threshold, the value threshold may be updated to the target value to appropriately adjust the target value. In an embodiment of the present invention, the value threshold may be determined based on expert experience and / or market conditions; the value threshold may be adjusted in response to a first threshold adjustment instruction; and the value threshold may be adjusted based on market changes.

[0048] In an embodiment of the present invention, after determining the object value of the target object based on the first object data and the target probability, a pre-set value correction rule can also be obtained. The value correction rule can be determined by an expert through business knowledge and experience, so as to correct or supplement the object value through the value correction rule, so that the expert can participate in the object value determination with his own experience, thereby affecting business decisions; according to the value correction rule, the object value is corrected, and the object value is updated according to the correction result obtained. For example, the object value indicates that the target object may be lost, but according to the value correction rule, it is determined that the target object is a long-term major customer of the company, then the object value can be modified.

[0049] In an embodiment of the present invention, after determining the value of an object, the object value level can be accurately identified based on the object value, such as whether the target object is a high-value object. For example, if the object value is greater than a preset value threshold, the target object is regarded as a high-value object. According to the object value level, personalized marketing strategies and service measures for the target object are determined, so as to implement differentiated marketing strategies for objects of different value levels, such as providing exclusive discounts or conducting more intensive marketing in the next step for high-value objects, and providing retention measures for low-value objects that are potentially lost, thereby improving the satisfaction and loyalty of the target object, optimizing resource allocation, reducing ineffective marketing, and improving overall operational efficiency. The risk level of the target object can also be evaluated based on the object value, so as to specify reasonable credit policies and other business processing policies and loan interest rates and other business processing contents according to the risk level. The above technical solution can not only achieve efficient and accurate determination of the object value, but also implement a series of refined operation strategies based on the shortcomings of the object value, thereby gaining advantages in fierce market competition and achieving sustainable growth in the scale of business processing.

[0050] The solution of the embodiment of the present invention determines the object value based on the first object data and the business processing prediction model. It can not only achieve efficient and accurate determination of the object value, but also make in-depth application of existing data, continuously improve the analysis and processing capabilities, focus on value analysis, realize the centralized collection and processing of object due diligence, and provide high-quality object value judgment services.

[0051] The technical solution of the embodiment of the present invention obtains first object data, and predicts the target probability of the target object handling the target business based on the first object data and the business handling prediction model, so as to facilitate the subsequent determination of the object value through the target probability, wherein the first object data includes the data of the target object under the associated features associated with the target business, and then the object value of the target object is determined based on the first object data and the target probability to achieve the determination of the object value, wherein the business handling prediction model is obtained by training based on at least one associated sample, and the associated sample includes the data of the sample object under the associated features. The above technical solution does not need to rely on manual labor, and the business prediction model obtained by training the associated samples including the data under the associated features associated with the target business can make the business prediction model more targeted and automatically predict the probability of handling the target business, thereby efficiently and accurately predicting the target probability, and then achieving efficient and accurate determination of the object value based on the target probability.

[0052] An optional technical solution, after predicting the target probability of the target object handling the target business based on the first object data and the business handling prediction model, the object value determination method also includes: obtaining a probability threshold, and when the target probability is greater than the probability threshold, updating the probability threshold to the target probability.

[0053] The probability threshold can be understood as the maximum value of the target probability when the probability threshold is not updated to the target probability.

[0054] In the implementation of the present invention, for example, when verifying the business processing prediction model through a validation set, the probability threshold can be determined by using a Receiver Operating Characteristic Curve (ROC) curve or a Precision-Recall Curve (PR) curve; for another example, when verifying the business processing prediction model through a validation set, the probability threshold can be obtained by weighing at least one of the indicators such as the true positive rate, the false positive rate, the precision rate and the recall rate.

[0055] In the embodiment of the present invention, a probability threshold is obtained, and when the target probability is greater than the probability threshold, the probability threshold is updated to the target probability, which can improve the accuracy of the target probability.

[0056] Based on the above scheme, another optional technical scheme, after the business processing prediction model is trained, the object value determination method also includes: obtaining the data to be predicted and the label corresponding to the data to be predicted, wherein the data to be predicted includes the data of the object to be predicted under the associated features, and the label represents whether the object to be predicted has handled the target business; inputting the data to be predicted into the business processing prediction model to obtain the prediction result output by the business processing prediction model, and using the prediction result and the label as a set of target data; based on multiple sets of target data, determining the true positive rate and false positive rate of the business processing prediction model; and determining the probability threshold according to the true positive rate and the false positive rate.

[0057] The data to be predicted can be understood as data used to be input into the business processing prediction model so that the business processing prediction model predicts the probability of the object to be predicted handling the target business; the data to be predicted may include data of the object to be predicted under associated features.

[0058] The label can be understood as the label corresponding to the data to be predicted, and the label can represent whether the object to be predicted has handled the target business; the data to be predicted and the label corresponding to the data to be predicted can be understood as a set of verification samples.

[0059] The object to be predicted can be understood as the object corresponding to the data to be predicted.

[0060] In the embodiment of the present invention, data to be predicted and labels corresponding to the data to be predicted may be obtained.

[0061] The prediction result can be understood as the result of the prediction model predicting the probability of the predicted object handling the target business.

[0062] The target data can be understood as data including the prediction results and the labels corresponding to the data to be predicted corresponding to the prediction results.

[0063] In an embodiment of the present invention, the data to be predicted is input into a business processing prediction model to obtain a prediction result, and the prediction result and the label are used as a set of target data.

[0064] The True Positive Rate (TPR) and the False Positive Rate (FPR) are both core indicators for measuring the model's ability to classify positive and negative samples in classification model evaluation. TPR represents the proportion of positive examples correctly predicted by the model to all actual positive examples, reflecting the model's ability to identify positive examples. FPR represents the proportion of negative examples incorrectly predicted as positive by the model to all actual negative examples, reflecting the degree of misjudgment of negative examples by the model. The True Positive Rate can also be called the Positive Prediction Rate; the False Positive Rate can also be called the Negative Prediction Rate.

[0065] In the embodiment of the present invention, the true positive rate and the false positive rate may be determined based on multiple sets of target data, and the probability threshold may be determined according to the true positive rate and the false positive rate.

[0066] For example, the probability threshold BestThreshold can be determined according to the true positive rate and the false positive rate by the formula BestThreshold=argmax(TPR-FPR), that is, the difference between the true positive rate and the false positive rate when the difference is the largest can be used as the probability threshold.

[0067] In the embodiment of the present invention, after the probability threshold is determined, the probability threshold may be adjusted in response to a second threshold adjustment instruction; the probability threshold may also be adjusted according to market changes.

[0068] In an embodiment of the present invention, the true positive rate and the false positive rate are determined, and the probability threshold is determined based on the true positive rate and the false positive rate, so that the probability threshold can be more closely matched with the business processing prediction model, thereby improving the accuracy of the probability threshold.

[0069] Another optional technical solution is that after predicting the target probability of the target object handling the target business based on the first object data and the business handling prediction model, the object value determination method also includes: obtaining a pre-set probability correction rule; performing probability correction on the target probability according to the probability correction rule, and updating the target probability based on the obtained correction result.

[0070] Among them, the probability correction rule can be understood as a rule for correcting the target probability.

[0071] The correction result can be understood as the result obtained by probability correction of the target probability according to the probability correction rule.

[0072] The solution of the embodiment of the present invention obtains probability correction rules, performs probability correction on the target probability according to the probability correction rules, and updates the target probability based on the obtained correction results. This can realize the combined use of the business processing prediction model and the probability correction rules, thereby improving the accuracy and reliability of the target probability, and further improving the accuracy of the object value determination.

[0073] Figure 2 This is a flow chart of another object value determination method provided in an embodiment of the present invention. This embodiment is an optimization based on the above-mentioned technical solutions. In this embodiment, the associated features are optionally determined by the following steps: obtaining historical data and determining at least one candidate feature based on the historical data; and determining associated features from the at least one candidate feature based on the target business. Explanations of terms that are identical or corresponding to those in the above-mentioned embodiments are not repeated here.

[0074] See also Figure 2 The method of this embodiment may specifically include the following steps:

[0075] S210: Acquire historical data, and determine at least one feature to be selected based on the historical data.

[0076] Among them, historical data can be understood as data generated historically; historical data can include second object data of historical objects that have handled target businesses, and can also include data corresponding to non-target businesses; historical data can, for example, include at least one of the object feedback opinions received by the business processing system, various historical business processing data stored, and second object data, etc.

[0077] In an embodiment of the present invention, historical data may be acquired, and at least one feature to be selected may be determined based on the historical data. For example, feature extraction may be performed on the historical data to determine at least one feature to be selected.

[0078] In an embodiment of the present invention, after acquiring historical data, data cleaning may be performed on the historical data to remove abnormal values and missing values in the historical data, and the historical data may be updated based on the obtained data cleaning results to ensure the quality of the historical data.

[0079] Optionally, the historical data includes second object data for historical objects that have handled the target business. This can increase the proportion of associated features in the at least one selected feature determined based on the historical data, thereby improving the efficiency of subsequently determining associated features from the at least one selected feature. A historical object is an object that has historically handled the target business. The second object data can be understood as data related to the historical object's handling of the target business.

[0080] S220: Determine a related feature from at least one feature to be selected according to the target business.

[0081] In the embodiment of the present invention, the associated feature may be determined from at least one feature to be selected according to the target service.

[0082] Exemplarily, at least one of the selected features may include loan amount, loan term, credit rating, loan interest rate, age of the borrower, installment amount, loan grade, sub-grade of loan grade, employment title, years of employment, house ownership status provided by the borrower at the time of registration, annual income, purpose of loan and repayment period. Based on the target business, the associated features determined from at least one of the selected features are loan amount, purpose of loan, annual income, credit rating and repayment period.

[0083] In an embodiment of the present invention, considering that wide tables are often used to integrate information from multiple data sources, they can merge information from multiple data sources into a single table containing all relevant variables to facilitate data analysis and statistical modeling. Since all relevant information in a wide table is stored in the same row, complex table join operations are not required. Therefore, wide tables can make data visualization and exploratory analysis more intuitive and direct. The structure of wide tables can also make data query and extraction relatively simple. On this basis, the number of associated features can be at least one. After determining that at least one associated feature is obtained, the at least one associated feature can be placed in the wide table as a necessary field of the wide table to obtain a wide table template that integrates the at least one associated feature in the form of a wide table. Subsequently, the data under the associated features of the target object associated with the target business can be filled into the wide table template to obtain a target wide table, which is used as the first object data. It should be noted that the associated samples used to train the business processing prediction model can also be represented in the form of a wide table. That is, the data under the associated features of the sample object can be filled into the wide table template to obtain associated samples to facilitate model training.

[0084] For example, see Figure 3 , it is possible to collect historical data including historical object data of multiple historical objects; perform outlier processing on the historical data; extract characteristic elements, related tables and related fields from the historical data to determine at least one candidate feature; perform feature screening on at least one candidate feature according to the target business to screen out associated features; perform feature deriving on the associated features; encode the associated features into a form suitable for model processing; and obtain a wide table template based on the associated features.

[0085] S230. Obtain first object data, and predict the target probability of the target object handling the target business based on the first object data and the business handling prediction model, wherein the first object data includes data of the target object under associated features associated with the target business, and the business handling prediction model is trained based on at least one associated sample, and the associated sample includes data of the sample object under the associated features.

[0086] S240: Determine the object value of the target object according to the first object data and the target probability.

[0087] The technical solution of the embodiment of the present invention can improve the comprehensiveness of the determined associated features by obtaining historical data, determining at least one candidate feature based on the historical data, and then determining associated features from the at least one candidate feature based on the target business.

[0088] An optional technical solution, according to the target business, determines the associated features from at least one candidate feature, including: for each of the at least one candidate feature, determining feature data from historical data according to the candidate feature, and determining the feature importance of the candidate feature compared to the target business based on the feature data, wherein the feature data includes data belonging to the candidate feature; and determining the associated features from at least one candidate feature according to the feature importance corresponding to the at least one candidate feature.

[0089] Among them, feature data can be understood as data belonging to the selected feature; feature data can be, for example, data of a business transaction belonging to the selected feature, and feature data can be, for example, data related to a historical object belonging to the selected feature, and so on.

[0090] Feature importance can be used to understand the importance of the selected feature compared to the target business.

[0091] In the embodiment of the present invention, for each candidate feature, feature data may be determined from historical data based on the candidate feature, and the feature importance of the candidate feature compared to the target business may be determined based on the feature data.

[0092] In an embodiment of the present invention, an associated feature can be determined from at least one candidate feature based on the feature importance corresponding to the at least one candidate feature. For example, an associated feature can be determined from at least one candidate feature based on an importance threshold and the feature importance corresponding to the at least one candidate feature. Specifically, for example, a candidate feature with a feature importance less than the importance threshold and little relevance to the target business can be discarded, and the remaining candidate features can be used as associated features.

[0093] In an embodiment of the present invention, the feature importance corresponding to at least one candidate feature is determined, and an associated feature is determined from the at least one candidate feature based on the feature importance corresponding to the at least one candidate feature, so as to improve the accuracy of the associated feature.

[0094] Based on the above scheme, another optional technical scheme is that the number of feature data is at least one; based on the feature data, the feature importance of the feature to be selected compared to the target business is determined, including: for each feature data in the at least one feature data, based on the feature data, determining the information gain of the feature to be selected compared to the feature data under the target business; based on the information gain corresponding to the at least one feature data, determining the feature importance of the feature to be selected compared to the target business.

[0095] It should be noted that information gain is an important concept used for feature selection in the decision tree algorithm. It is based on the concept of entropy in information theory. Information gain measures the extent to which a feature can reduce the uncertainty of a data set (i.e., entropy). When constructing a decision tree, the feature with the largest information gain is usually selected as the node for splitting. Taking the above situation into consideration, for each feature data, based on the feature data, the information gain of the candidate feature compared to the feature data under the target business can be determined, and then the information gain corresponding to at least one feature data can be determined to determine the feature importance of the candidate feature compared to the target business.

[0096] For example, the information gain corresponding to at least one feature data can be calculated by the formula Determine the feature importance, where Importance(f) is the feature importance, N is the number of feature data, (f, x i ) is the feature f to be selected compared to x in the target business i The information gain of

[0097] In an embodiment of the present invention, the feature importance of the selected feature compared to the target business is determined based on the information gain corresponding to at least one feature data, which can ensure that the feature importance is the importance of the selected feature compared to the target business, thereby improving the accuracy of the determined feature importance.

[0098] Another optional technical solution is to determine at least one feature to be selected based on historical data, including: performing feature extraction processing on the historical data to obtain at least one first feature; and / or performing feature derivation processing on the historical data to obtain at least one second feature; and determining at least one feature to be selected based on the at least one first feature and / or the at least one second feature.

[0099] The first feature can be understood as a feature obtained by performing feature extraction processing on historical data.

[0100] In an embodiment of the present invention, feature extraction processing can be performed on historical data to obtain at least one first feature. For example, feature extraction processing can be performed on historical data to extract effective features that can be used for model training of the data thereunder as the first feature to obtain at least one candidate feature.

[0101] The second feature can be understood as a feature obtained by performing feature derivation processing on historical data.

[0102] In an embodiment of the present invention, feature derivation processing can be performed on historical data to obtain at least one second feature. For example, a derivative processing method can be determined to derive various data in the historical data. Based on the derivative processing method, at least one second feature is determined. For example, if there are multiple monthly consumption amount data in the historical data, the multiple monthly consumption amount data can be averaged, that is, the derivative processing method is average processing. In this case, a second feature can be determined as the monthly average consumption amount.

[0103] In the embodiment of the present invention, at least one feature to be selected may be determined based on at least one first feature and / or at least one second feature. For example, at least one first feature and / or at least one second feature may be used as at least one feature to be selected.

[0104] In the embodiment of the present invention, by performing feature extraction processing and / or feature derivation processing on historical data to determine at least one feature to be selected, the comprehensiveness of the determined at least one feature to be selected can be improved.

[0105] Another optional technical solution, the object value determination method also includes: in response to the to-be-selected feature supplementation instruction, obtaining the supplementary features associated with the target business; after determining at least one to-be-selected feature based on historical data, the object value determination method also includes: adding the supplementary features to at least one to-be-selected feature.

[0106] The feature supplementation instruction to be selected may be understood as an instruction to obtain supplementary features.

[0107] A supplementary feature can be understood as a feature that needs to be supplemented with at least one candidate feature.

[0108] In an embodiment of the present invention, supplementary features can be obtained in response to an instruction to supplement features to be selected. For example, supplementary features supplemented by experts based on their experience can be obtained in response to an instruction to supplement features to be selected, so that the experience of experts can be applied to the scenario of determining the value of the object, thereby improving the accuracy of the target probability predicted by the business processing prediction model; the supplementary features are added to at least one feature to be selected.

[0109] Exemplarily, in response to an instruction to supplement features to be selected, supplementary features including the month of issuance of the object, the category of loan purpose of the borrower when applying for the loan, and the frequency of loan application can be obtained; and the above supplementary features are added to at least one feature to be selected.

[0110] In an embodiment of the present invention, by responding to a candidate feature supplement instruction, obtaining supplementary features associated with the target service, and adding the supplementary features to at least one candidate feature, the comprehensiveness of the determined at least one candidate feature can be improved.

[0111] In order to better understand the technical solution of the above embodiment of the present invention, an optional example is provided here. Figure 4 Determining object value can include the following four steps: feature engineering, creating a wide table, XGBoost model training, probabilistic correction rule decision-making, and object value scoring. Feature engineering may include analyzing historical data content and extracting feature elements to determine at least one candidate feature, determining associated features from the at least one candidate feature based on the target business, and encoding the associated features into a format suitable for model processing. Creating a wide table may include creating a wide table template based on the associated features, including necessary fields such as asset number and type, currency type, and number of days overdue; and populating the wide table template to obtain at least one associated sample. XGBoost model training may include developing and training the XGBoost model. Probabilistic correction rule decision-making may include determining probabilistic correction rules, including rules for distinguishing existing and new objects, identifying high-quality objects with more than 10 transactions, and activating moderately active customers. Object value scoring may include running the XGBoost model and performing object value scoring based on the probabilistic correction rules and the XGBoost model.

[0112] Figure 5 This is a block diagram of the structure of the object value determination device provided in an embodiment of the present invention. The device is used to execute the object value determination method provided in any of the above embodiments. The device and the object value determination method of the above embodiments belong to the same inventive concept. For details not fully described in the embodiment of the object value determination device, please refer to the embodiment of the above object value determination method. Figure 5 , the device may specifically include: a target probability prediction module 310 and an object value determination module 320.

[0113] The target probability prediction module 310 is configured to obtain first object data and, based on the first object data and the business handling prediction model, predict the target probability of the target object handling the target business, wherein the first object data includes data of the target object under the associated features associated with the target business;

[0114] an object value determination module 320 for determining an object value of a target object based on the first object data and the target probability;

[0115] The business processing prediction model is trained based on at least one associated sample, and the associated sample includes data of the sample object under associated features.

[0116] Optionally, the device may further include the following modules to determine the associated features:

[0117] A candidate feature determination module is used to obtain historical data and determine at least one candidate feature based on the historical data;

[0118] The associated feature determination module is used to determine an associated feature from at least one feature to be selected according to the target business.

[0119] Optionally, based on the above device, the correlation feature determination module may include:

[0120] a feature importance determination submodule, configured to determine, for each of the at least one candidate feature, feature data from historical data based on the candidate feature, and determine, based on the feature data, a feature importance of the candidate feature relative to the target business, wherein the feature data includes data belonging to the candidate feature;

[0121] The associated feature determination submodule is used to determine an associated feature from at least one candidate feature according to the feature importance corresponding to the at least one candidate feature.

[0122] Optionally, based on the above device, the number of characteristic data is at least one;

[0123] The feature importance determination submodule may include:

[0124] an information gain determining unit, configured to determine, for each feature data of the at least one feature data, an information gain of the candidate feature compared to the feature data under the target service according to the feature data;

[0125] The feature importance determination unit is used to determine the feature importance of the candidate feature compared to the target business based on the information gain corresponding to at least one feature data.

[0126] Optionally, based on the above device, the module for determining the feature to be selected may include:

[0127] A first feature obtaining submodule is configured to perform feature extraction processing on historical data to obtain at least one first feature; and / or,

[0128] A second feature obtaining submodule is used to perform feature derivation processing on historical data to obtain at least one second feature;

[0129] The feature-to-be-selected determining submodule is configured to determine at least one feature to be selected based on at least one first feature and / or at least one second feature.

[0130] Optionally, based on the above device, the device may further include:

[0131] A supplementary feature acquisition module, configured to acquire, in response to a feature supplement instruction, supplementary features associated with a target service;

[0132] The supplementary feature supplementation module is used to supplement the supplementary feature into the at least one feature to be selected after determining at least one feature to be selected based on historical data.

[0133] Optionally, the historical data includes second object data of historical objects that have handled the target business.

[0134] Optionally, the device may further include:

[0135] The target probability update module is used to obtain a probability threshold after predicting the target probability of the target object handling the target business based on the first object data and the business handling prediction model, and update the probability threshold to the target probability when the target probability is greater than the probability threshold.

[0136] Optionally, based on the above device, the device may further include:

[0137] The label acquisition module is used to obtain the data to be predicted and the labels corresponding to the data to be predicted after training the business processing prediction model. The data to be predicted includes the data of the object to be predicted under the associated features, and the labels indicate whether the object to be predicted has processed the target business;

[0138] The target data is used as a module to input the data to be predicted into the business processing prediction model, obtain the prediction results output by the business processing prediction model, and use the prediction results and labels as a set of target data;

[0139] A false positive rate determination module is used to determine the true positive rate and false positive rate of the business processing prediction model based on multiple sets of target data;

[0140] The probability threshold determination module is used to determine the probability threshold according to the true positive rate and the false positive rate.

[0141] Optionally, the device may further include:

[0142] A probability correction rule acquisition module is used to obtain a preset probability correction rule after predicting the target probability of the target object handling the target business based on the first object data and the business handling prediction model;

[0143] The target probability updating module is used to perform probability correction on the target probability according to the probability correction rule, and update the target probability according to the obtained correction result.

[0144] The object value determination device provided by the embodiment of the present invention obtains the first object data through the target probability prediction module, and predicts the target probability of the target object handling the target business based on the first object data and the business handling prediction model, so as to facilitate the subsequent determination of the object value through the target probability, wherein the first object data includes the data of the target object under the associated features associated with the target business; and then determines the object value of the target object based on the first object data and the target probability through the object value determination module to achieve the determination of the object value, wherein the business handling prediction model is obtained by training based on at least one associated sample, and the associated sample includes the data of the sample object under the associated features. The above-mentioned device, without relying on manual labor, can make the business prediction model automatically predict the probability of handling the target business more targeted, thereby efficiently and accurately predicting the target probability, and then efficiently and accurately determining the object value.

[0145] The object value determination device provided in the embodiment of the present invention can execute the object value determination method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0146] It is worth noting that in the embodiment of the above-mentioned object value determination device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the various functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0147] Figure 6 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0148] like Figure 6As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., which is communicatively connected to the at least one processor 11. The memory stores a computer program that can be executed by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the RAM 13. The processor 11, ROM 12, and RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0149] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0150] The processor 11 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors that run machine learning model algorithms, a digital signal processor (DSP), and any other suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the object value determination method.

[0151] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product that includes a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication unit 19, or installed from the storage unit 18, or installed from the ROM 12. When the computer program is executed by the processor 11, the above-mentioned functions defined in the method of the embodiment of the present invention are performed.

[0152] In some embodiments, the object value determination method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the object value determination method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the object value determination method in any other suitable manner (e.g., via firmware).

[0153] Various embodiments of the systems and techniques described above can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on a chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0154] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0155] In the context of the present invention, computer-readable storage media can be tangible media that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage media can be machine-readable signal media. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0156] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0157] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0158] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0159] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0160] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A method for determining the value of an object, characterized in that: include: Acquire first object data, and predict a target probability of a target object handling a target business based on the first object data and a business handling prediction model, wherein the first object data includes data of the target object under associated features associated with the target business; determining an object value of the target object based on the first object data and the target probability; The business processing prediction model is trained based on at least one associated sample, and the associated sample includes data of the sample object under the associated feature.

2. The method according to claim 1, characterized in that The associated features are determined by the following steps: Acquiring historical data, and determining at least one feature to be selected based on the historical data; The associated feature is determined from the at least one feature to be selected according to the target business.

3. The method according to claim 2, characterized in that The determining, based on the target business, the associated feature from the at least one feature to be selected includes: For each of the at least one candidate feature, determining feature data from the historical data based on the candidate feature, and determining feature importance of the candidate feature compared to the target business based on the feature data, wherein the feature data includes data belonging to the candidate feature; The associated feature is determined from the at least one feature to be selected according to the feature importance respectively corresponding to the at least one feature to be selected.

4. The method according to claim 3, characterized in that The number of the feature data is at least one; The determining, based on the feature data, the feature importance of the candidate feature compared to the feature importance of the target business includes: For each feature data of the at least one feature data, determining, based on the feature data, an information gain of the candidate feature compared to the feature data under the target service; Determine the feature importance of the candidate feature compared to the target business according to the information gain corresponding to the at least one feature data.

5. The method according to claim 2, characterized in that The determining, based on the historical data, at least one feature to be selected includes: Performing feature extraction processing on the historical data to obtain at least one first feature; and / or, Performing feature derivation processing on the historical data to obtain at least one second feature; At least one feature to be selected is determined according to the at least one first feature and / or the at least one second feature.

6. The method according to claim 2, characterized in that Also includes: In response to the feature supplement instruction, obtaining the supplementary feature associated with the target service; After determining at least one feature to be selected based on the historical data, the method further includes: The supplementary feature is added to the at least one feature to be selected.

7. The method according to claim 2, characterized in that The historical data includes second object data of historical objects that have handled the target business.

8. The method according to claim 1, characterized in that After predicting the target probability of the target object handling the target business based on the first object data and the business handling prediction model, the method further includes: A probability threshold is obtained, and when the target probability is greater than the probability threshold, the probability threshold is updated to the target probability.

9. The method according to claim 8, characterized in that After the business processing prediction model is obtained through training, the method further includes: Acquire data to be predicted and a label corresponding to the data to be predicted, wherein the data to be predicted includes data of the object to be predicted under the associated feature, and the label indicates whether the object to be predicted has handled the target business; Inputting the data to be predicted into the business processing prediction model to obtain a prediction result output by the business processing prediction model, and using the prediction result and the label as a set of target data; Determining a true positive rate and a false positive rate of the business processing prediction model based on the multiple sets of target data; The probability threshold is determined according to the true positive rate and the false positive rate.

10. The method according to claim 1, characterized in that After predicting the target probability of the target object handling the target business based on the first object data and the business handling prediction model, the method further includes: Obtaining pre-set probability correction rules; According to the probability correction rule, the target probability is probability corrected, and according to the obtained correction result, the target probability is updated.

Images