Method for updating object evaluation model, object evaluation method and device

Abstract

The present disclosure relates to an object evaluation model updating method, object evaluation method and device, and belongs to the field of financial technology, information technology and artificial intelligence. The method comprises: obtaining a first data set and inputting the first data set into an object evaluation model to output a first evaluation result set from the object evaluation model; obtaining a second data set updated from the first data set and inputting the second data set into the object evaluation model to output a second evaluation result set from the object evaluation model; judging whether a model updating condition is met based on the first evaluation result set and the second evaluation result set; and updating the object evaluation model if the model updating condition is met. Thus, automatic judgment of the model updating condition can be realized, and continuous updating of the model can be realized in the use process of the object evaluation model, the precision of the object evaluation model is improved, and the method is suitable for updating scenarios of user credit evaluation models.

Description

Technical Field

[0001] This disclosure relates to the fields of financial technology, information technology industry, and artificial intelligence technology, and in particular to a method for updating an object evaluation model, an object evaluation method, an apparatus, an electronic device, a computer-readable storage medium, and a computer program product. Background Technology

[0002] Currently, with the continuous development of artificial intelligence technology, models are widely used in the evaluation field, offering advantages such as high automation and low labor costs. For example, object feature data can be input into an object evaluation model, which then outputs the object's evaluation result. However, object evaluation models in related technologies suffer from low accuracy. Summary of the Invention

[0003] This disclosure provides a method for updating an object evaluation model, an object evaluation method, an apparatus, an electronic device, a computer-readable storage medium, and a computer program product, to at least solve the problem of low accuracy in object evaluation models in related technologies. The technical solution of this disclosure is as follows:

[0004] According to a first aspect of the present disclosure, a method for updating an object evaluation model is provided, comprising: acquiring a first dataset and inputting the first dataset into the object evaluation model, and having the object evaluation model output a first evaluation result set; acquiring a second dataset updated from the first dataset and inputting the second dataset into the object evaluation model, and having the object evaluation model output a second evaluation result set; determining, based on the first evaluation result set and the second evaluation result set, whether a model update condition is met; and updating the object evaluation model if the model update condition is met.

[0005] In one embodiment of this disclosure, determining whether the model update condition is met based on the first evaluation result set and the second evaluation result set includes: determining whether the model update condition is met based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set.

[0006] In one embodiment of this disclosure, determining whether the model update condition is met based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set includes: obtaining a first indicator based on the first dataset and the second dataset; obtaining a second indicator based on the first evaluation result set and the second evaluation result set; obtaining a third indicator and a fourth indicator based on the first evaluation result set; and determining whether the model update condition is met based on at least one of the first indicator, the second indicator, the third indicator, and the fourth indicator.

[0007] In one embodiment of this disclosure, the first dataset and the second dataset include category dimensions to which the objects belong. The step of obtaining a first indicator based on the first dataset and the second dataset includes: obtaining a first number of objects under each category dimension in the first dataset; obtaining a second number of objects under each category dimension in the second dataset; and obtaining the first indicator based on the first number and the second number.

[0008] In one embodiment of this disclosure, obtaining the first index based on the first quantity and the second quantity includes: obtaining a fifth index under the i-th category dimension based on the first quantity and the second quantity under the i-th category dimension; and obtaining the first index based on the average of the fifth index under N category dimensions; wherein i is a positive integer not greater than N, and N is a positive integer.

[0009] In one embodiment of this disclosure, the first evaluation result set and the second evaluation result set include the level dimension to which the object belongs. The step of obtaining the second indicator based on the first evaluation result set and the second evaluation result set includes: obtaining a third number of objects under each level dimension in the first evaluation result set; obtaining a fourth number of objects under each level dimension in the second evaluation result set; and obtaining the second indicator based on the third number and the fourth number.

[0010] In one embodiment of this disclosure, obtaining the second index based on the third quantity and the fourth quantity includes: obtaining a sixth index under the i-th category dimension based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension; obtaining a seventh index under the i-th category dimension based on the variance of the fourth quantity under multiple level dimensions belonging to the i-th category dimension; and obtaining the second index based on the variance of the sixth index under N category dimensions and the variance of the seventh index under N category dimensions; wherein i is a positive integer not greater than N, and N is a positive integer.

[0011] In one embodiment of this disclosure, the first evaluation result set includes the level dimension to which the object belongs. The step of obtaining the third and fourth indicators based on the first evaluation result set includes: obtaining the third number of objects under each level dimension in the first evaluation result set; obtaining a sixth indicator under the i-th category dimension based on the variance of the third number under multiple level dimensions belonging to the i-th category dimension; obtaining the third indicator based on the skewness of the sixth indicator under N category dimensions; and obtaining the fourth indicator based on the kurtosis of the sixth indicator under N category dimensions; wherein i is a positive integer not greater than N, and N is a positive integer.

[0012] In one embodiment of this disclosure, the first dataset and the second dataset include category dimensions to which the object belongs, and the first evaluation result set includes level dimensions to which the object belongs; wherein, updating the object evaluation model includes: obtaining a fifth indicator under N category dimensions based on the first dataset and the second dataset, where N is a positive integer; obtaining a sixth indicator under N category dimensions based on the first evaluation result set; selecting a target category dimension from the N category dimensions based on the fifth indicator and the sixth indicator under N category dimensions; determining a first model parameter associated with the target category dimension from the model parameters of the object evaluation model; and updating the first model parameter.

[0013] In one embodiment of this disclosure, obtaining the fifth indicator under N category dimensions based on the first dataset and the second dataset includes: obtaining a first number of objects under each category dimension in the first dataset; obtaining a second number of objects under each category dimension in the second dataset; and obtaining the fifth indicator under the i-th category dimension based on the first number and the second number under the i-th category dimension, where i is a positive integer not greater than N.

[0014] In one embodiment of this disclosure, obtaining the sixth index under N category dimensions based on the first evaluation result set includes: obtaining the third number of objects under each level dimension in the first evaluation result set; and obtaining the sixth index under the i-th category dimension based on the variance of the third number under multiple level dimensions belonging to the i-th category dimension, where i is a positive integer not greater than N.

[0015] In one embodiment of this disclosure, updating the first model parameters includes: obtaining model parameters of at least one set model; and performing a weighted summation of the first model parameters and the model parameters of at least one set model to obtain the second model parameters after updating the first model parameters.

[0016] According to a second aspect of the present disclosure, an object evaluation method is provided, comprising: acquiring target data of an object; inputting the target data into an object evaluation model, and outputting an evaluation result of the object by the object evaluation model, wherein the object evaluation model is obtained using the object evaluation model update method as described in the first aspect above.

[0017] According to a third aspect of the present disclosure, an apparatus for updating an object evaluation model is provided, comprising: a first acquisition module configured to acquire a first dataset and input the first dataset into the object evaluation model, thereby outputting a first evaluation result set by the object evaluation model; a second acquisition module configured to acquire a second dataset updated from the first dataset and input the second dataset into the object evaluation model, thereby outputting a second evaluation result set by the object evaluation model; a judgment module configured to determine, based on the first evaluation result set and the second evaluation result set, whether a model update condition is met; and an update module configured to update the object evaluation model if the model update condition is met.

[0018] In one embodiment of this disclosure, the judgment module is further configured to: determine whether the model update condition is met based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set.

[0019] In one embodiment of this disclosure, the judgment module is further configured to: obtain a first indicator based on the first dataset and the second dataset; obtain a second indicator based on the first evaluation result set and the second evaluation result set; obtain a third indicator and a fourth indicator based on the first evaluation result set; and determine whether the model update condition is met based on at least one of the first indicator, the second indicator, the third indicator, and the fourth indicator.

[0020] In one embodiment of this disclosure, the first dataset and the second dataset include category dimensions to which the objects belong, and the judgment module is further configured to: obtain a first number of objects under each category dimension in the first dataset; obtain a second number of objects under each category dimension in the second dataset; and obtain the first index based on the first number and the second number.

[0021] In one embodiment of this disclosure, the judgment module is further configured to: obtain a fifth indicator under the i-th category dimension based on the first quantity and the second quantity under the i-th category dimension; and obtain the first indicator based on the average of the fifth indicators under N category dimensions; wherein i is a positive integer not greater than N, and N is a positive integer.

[0022] In one embodiment of this disclosure, the first evaluation result set and the second evaluation result set include the level dimension to which the object belongs. The judgment module is further configured to: obtain a third number of objects under each level dimension in the first evaluation result set; obtain a fourth number of objects under each level dimension in the second evaluation result set; and obtain the second indicator based on the third number and the fourth number.

[0023] In one embodiment of this disclosure, the judgment module is further configured to: obtain a sixth index under the i-th category dimension based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension; obtain a seventh index under the i-th category dimension based on the variance of the fourth quantity under multiple level dimensions belonging to the i-th category dimension; and obtain a second index based on the variance of the sixth index under N category dimensions and the variance of the seventh index under N category dimensions; wherein i is a positive integer not greater than N, and N is a positive integer.

[0024] In one embodiment of this disclosure, the first evaluation result set includes the level dimension to which the object belongs, and the judgment module is further configured to: obtain the third quantity of objects under each level dimension in the first evaluation result set; obtain the sixth index under the i-th category dimension based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension; obtain the third index based on the skewness of the sixth index under N category dimensions; and obtain the fourth index based on the kurtosis of the sixth index under N category dimensions; wherein i is a positive integer not greater than N, and N is a positive integer.

[0025] In one embodiment of this disclosure, the first dataset and the second dataset include the category dimension to which the object belongs, and the first evaluation result set includes the level dimension to which the object belongs; wherein, the update module is further configured to: obtain a fifth indicator under N category dimensions based on the first dataset and the second dataset, where N is a positive integer; obtain a sixth indicator under N category dimensions based on the first evaluation result set; filter out a target category dimension from the N category dimensions based on the fifth indicator and the sixth indicator under N category dimensions; determine a first model parameter associated with the target category dimension from the model parameters of the object evaluation model; and update the first model parameter.

[0026] In one embodiment of this disclosure, the update module is further configured to: obtain a first number of objects under each category dimension in the first dataset; obtain a second number of objects under each category dimension in the second dataset; and obtain a fifth index under the i-th category dimension based on the first number and the second number under the i-th category dimension, where i is a positive integer not greater than N.

[0027] In one embodiment of this disclosure, the updating module is further configured to: obtain the third number of objects under each level dimension in the first evaluation result set; and obtain the sixth index under the i-th category dimension based on the variance of the third number under multiple level dimensions belonging to the i-th category dimension, where i is a positive integer not greater than N.

[0028] In one embodiment of this disclosure, the update module is further configured to: obtain model parameters of at least one set model; and perform a weighted summation of the first model parameters and the model parameters of at least one set model to obtain second model parameters updated from the first model parameters.

[0029] According to a fourth aspect of the present disclosure, an object evaluation apparatus is provided, comprising: an acquisition module configured to acquire target data of an object; and an evaluation module configured to input the target data into an object evaluation model and output an evaluation result of the object from the object evaluation model, wherein the object evaluation model is obtained using the object evaluation model update method as described in the first aspect above.

[0030] According to a fifth aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the object evaluation model update method as described in the first aspect above, and / or to implement the object evaluation method as described in the second aspect above.

[0031] According to a sixth aspect of the present disclosure, a computer-readable storage medium is provided, wherein when instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform an object evaluation model update method as described in the first aspect above, and / or to perform an object evaluation method as described in the second aspect above.

[0032] According to a seventh aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the object evaluation model update method as described in the first aspect above, and / or implements the object evaluation method as described in the second aspect above.

[0033] The technical solution provided by the embodiments of this disclosure brings at least the following beneficial effects: it can comprehensively consider the first evaluation result set and the second evaluation result set to determine whether the model update condition is met, so as to realize the automatic judgment of the model update condition, and update the object evaluation model when the model update condition is met, so as to realize the automatic update of the object evaluation model, that is, the model can be continuously updated during the use of the object evaluation model, improving the accuracy of the object evaluation model, and is suitable for the update scenario of user credit evaluation model.

[0034] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0035] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0036] Figure 1 This is a flowchart illustrating the method for updating an object evaluation model according to the first embodiment of this disclosure.

[0037] Figure 2 This is a flowchart illustrating the method for updating an object evaluation model according to a second embodiment of the present disclosure.

[0038] Figure 3 This is a flowchart illustrating the method for updating an object evaluation model according to a third embodiment of the present disclosure.

[0039] Figure 4 This is a schematic diagram of an update method for an object evaluation model according to the fourth embodiment of this disclosure.

[0040] Figure 5 This is a flowchart illustrating the object evaluation method according to the first embodiment of this disclosure.

[0041] Figure 6 This is a block diagram of an object evaluation model update apparatus according to a first embodiment of the present disclosure.

[0042] Figure 7 This is a block diagram of an object evaluation apparatus according to a first embodiment of the present disclosure.

[0043] Figure 8 This is a block diagram illustrating an electronic device according to an exemplary embodiment. Detailed Implementation

[0044] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0045] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0046] The acquisition, storage, use, and processing of data in this disclosed technical solution all comply with the relevant provisions of national laws and regulations.

[0047] Figure 1 This is a flowchart illustrating the method for updating an object evaluation model according to the first embodiment of this disclosure.

[0048] like Figure 1 As shown, the object evaluation model update method of the first embodiment of this disclosure includes the following steps:

[0049] S101, obtain the first dataset, input the first dataset into the object evaluation model, and the object evaluation model outputs the first evaluation result set.

[0050] It should be noted that the object evaluation model update method of this disclosure is executed by an electronic device. The object evaluation model update method of this disclosure can be executed by the object evaluation model update device of this disclosure, which can be configured in any electronic device to execute the object evaluation model update method of this disclosure.

[0051] It should be noted that the object evaluation model can be any object evaluation model in related technologies, without further limitations. For example, object evaluation models may include user credit evaluation models, product performance evaluation models, project risk evaluation models, etc. For example, the object evaluation model may be a Light GBM (Light Gradient Boosting Machine) model.

[0052] In the embodiments of this disclosure, the first dataset includes first data of multiple users in a user set, and the first evaluation result set includes first evaluation results of multiple users in the user set, with a one-to-one correspondence between the first data and the first evaluation result. For example, the user set includes users 1 to M, the first dataset includes the first data of user s, and the first evaluation result set includes the first evaluation result of user s, where s is a positive integer not greater than M, and M is a positive integer.

[0053] It should be noted that no excessive restrictions are placed on the first data or the first evaluation result.

[0054] For example, taking a user credit rating model as an example, the first data may include the user's characteristic data, and the first rating result may include the user's credit rating, credit score, etc. The user's characteristic data may include the user's income, age, years of employment, gender, industry category, etc.

[0055] For example, taking a product performance evaluation model as an example, the first data may include the product's characteristic data, and the first evaluation result may include the product's performance level, performance score, failure rate, etc. The product's characteristic data may include the product's materials, manufacturing process, appearance, etc.

[0056] For example, taking the object evaluation model as a risk assessment model for a project, the first data may include the project's characteristic data, and the first evaluation result may include the project's risk level, risk score, etc. The project's characteristic data may include the number of participants, project duration, project investment costs, and whether the project implementation area includes overseas regions, etc.

[0057] S102, obtain the second dataset after the first dataset is updated, and input the second dataset into the object evaluation model, and the object evaluation model outputs the second evaluation result set.

[0058] It should be noted that the relevant content of the second dataset and the second evaluation result set can be referred to the relevant content of the first dataset and the first evaluation result set in the above embodiments, and will not be repeated here.

[0059] It is understandable that the first dataset may contain inaccuracies, and the second dataset refers to the updated result of the first dataset.

[0060] In one implementation, obtaining the updated second dataset after the first dataset includes determining whether a dataset update condition is met; if the dataset update condition is met, the updated second dataset after the first dataset is obtained. Therefore, this method can obtain the updated second dataset only when the dataset update condition is met.

[0061] It should be noted that there are no strict restrictions on the conditions for updating the dataset. For example, it may include the difference between the time the first dataset was acquired and the current time being greater than or equal to a set threshold, or the current time reaching a set time.

[0062] S103, based on the first evaluation result set and the second evaluation result set, determine whether the model update conditions are met.

[0063] In the embodiments of this disclosure, based on the first evaluation result set and the second evaluation result set, it is determined whether the model update conditions are met, including the following possible implementation methods:

[0064] Method 1: Based on the first evaluation result set and the second evaluation result set, obtain the second indicator, and based on the second indicator, determine whether the model update conditions are met.

[0065] In some examples, the model update condition is determined based on the second indicator, including whether the second indicator is greater than or equal to a set threshold. If the second indicator is greater than or equal to the set threshold, the model update condition is determined to be met; if the second indicator is less than the set threshold, the model update condition is determined not to be met.

[0066] Method 2: Based on the first evaluation result set and the second evaluation result set, obtain the second indicator; based on the first evaluation result set, obtain the third indicator and the fourth indicator; based on at least one of the second indicator, the third indicator and the fourth indicator, determine whether the model update conditions are met.

[0067] In some examples, the model update conditions are determined based on at least one of the second, third, and fourth indicators, including obtaining the target indicator based on at least one of the second, third, and fourth indicators, and determining whether the model update conditions are met based on the target indicator.

[0068] For example, the calculation process for the target indicator is as follows:

[0069]

[0070] Where Sta is the target indicator, R is the second indicator, Sk is the third indicator, KT is the fourth indicator, and b, c, and d are coefficients.

[0071] For example, based on the target indicator, determine whether the model update conditions are met, including whether the target indicator is greater than or equal to a set threshold. If the target indicator is greater than or equal to the set threshold, it is determined that the model update conditions are not met; if the target indicator is less than the set threshold, it is determined that the model update conditions are met.

[0072] Method 3: Based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set, determine whether the model update conditions are met.

[0073] Therefore, this method can comprehensively consider the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set to determine whether the model update conditions are met.

[0074] In some examples, a target index is obtained based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set. Based on the target index, it is determined whether the model update conditions are met.

[0075] In some examples, the target indicator is obtained based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set. This includes obtaining multiple types of indicators based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set, and obtaining the target indicator based on multiple types of indicators.

[0076] It should be noted that the relevant content of the second, third, and fourth indicators can be found in the following examples, and will not be repeated here.

[0077] S104. If the model update conditions are met, update the object evaluation model.

[0078] It should be noted that model updates can be implemented using any model update method in the relevant technologies, and no further restrictions are imposed here.

[0079] In one implementation, updating the object evaluation model includes obtaining training samples, wherein the training samples include sample data and sample evaluation results, and updating the object evaluation model based on the training samples.

[0080] In one implementation, updating the object evaluation model includes obtaining model parameters of at least one set model, and performing a weighted summation of the model parameters of the object evaluation model and the model parameters of the at least one set model to obtain the updated model parameters of the object evaluation model.

[0081] It should be noted that there are no strict restrictions on the model setting. For example, it can include DNN (Deep Neural Network) models, LR (Logistic Regression) models, etc.

[0082] In one implementation, updating the object evaluation model includes obtaining at least one set model, combining the object evaluation model and the at least one set model to obtain an updated object evaluation model.

[0083] In summary, the object evaluation model update method provided in this embodiment involves obtaining a first dataset and inputting it into the object evaluation model, which then outputs a first evaluation result set. A second dataset updated from the first dataset is then obtained and input into the object evaluation model, which again outputs a second evaluation result set. Based on the first and second evaluation result sets, it is determined whether the model update conditions are met. If the conditions are met, the object evaluation model is updated. Therefore, by comprehensively considering both the first and second evaluation result sets, the method automatically determines whether the model update conditions are met, and updates the object evaluation model when these conditions are met. This automatic updating of the object evaluation model allows for continuous model updates during its use, improving its accuracy and making it suitable for updating user credit evaluation models.

[0084] Figure 2 This is a flowchart illustrating the method for updating an object evaluation model according to a second embodiment of the present disclosure.

[0085] like Figure 2 As shown, the object evaluation model update method of the second embodiment of this disclosure includes the following steps:

[0086] S201, Obtain the first dataset and input the first dataset into the object evaluation model, and the object evaluation model outputs the first evaluation result set.

[0087] S202, obtain the second dataset after updating the first dataset, and input the second dataset into the object evaluation model, and the object evaluation model outputs the second evaluation result set.

[0088] The details of steps S201-S202 can be found in the above embodiments and will not be repeated here.

[0089] S203, based on the first dataset and the second dataset, the first indicator is obtained.

[0090] In the embodiments of this disclosure, the first indicator is obtained based on the first dataset and the second dataset, and may include the following possible implementation methods:

[0091] Method 1: Obtain the first indicator based on data of the same dimension from the first dataset and the second dataset.

[0092] It should be noted that there are no excessive restrictions on the dimensions; for example, they can include industry category, years of work experience, etc.

[0093] In some examples, the first indicator is obtained based on data of the same dimension in the first dataset and the second dataset, including the average of multiple differences based on data of the same dimension in the first dataset and the second dataset.

[0094] Method 2: Obtain the first number of objects under each category dimension in the first dataset, and obtain the second number of objects under each category dimension in the second dataset. Based on the first and second numbers, obtain the first metric.

[0095] Therefore, this method can comprehensively consider the first number of objects under each category dimension in the first dataset and the second number of objects under each category dimension in the second dataset to obtain the first index.

[0096] It should be noted that the first and second datasets include the category dimension to which the objects belong. The category dimension may include industry category, gender, product category, etc.

[0097] It is understandable that the first quantity under different category dimensions may be different or the same, and the second quantity under different category dimensions may be different or the same; no further restrictions are imposed here.

[0098] In some examples, a first indicator is obtained based on a first quantity and a second quantity. This includes obtaining a fifth indicator for the i-th category dimension based on the first and second quantities under the i-th category dimension, and obtaining the first indicator based on the average of the fifth indicators across N category dimensions. Here, i is a positive integer not greater than N, and N is a positive integer. Thus, this method can obtain the fifth indicator for the same category dimension based on the first and second quantities, and obtain the first indicator based on the average of the fifth indicators across multiple category dimensions.

[0099] It should be noted that N is a positive integer not greater than the number of category dimensions. For example, N is the number of category dimensions.

[0100] In some examples, based on the first and second quantities under the i-th category dimension, a fifth indicator is obtained. This involves obtaining the absolute difference between the first and second quantities under the i-th category dimension, and using the ratio of the absolute difference to the first quantity as the fifth indicator under the i-th category dimension. For example, the calculation process for the fifth indicator under the i-th category dimension is as follows:

[0101]

[0102] Among them, C i X is the fifth indicator under the i-th category dimension. i1 X represents the first quantity under the i-th category dimension. i2This represents the second quantity under the i-th category dimension.

[0103] For example, taking the object evaluation model as a user credit evaluation model, the category dimension is the industry category to which the user belongs. If the first and second number of users under the sales industry category dimension are 100 and 95 respectively, then the fifth indicator under the sales industry category dimension is |100-95| / 100 = 5%. If the first and second number of users under the manufacturing industry category dimension are 80 and 90 respectively, then the fifth indicator under the manufacturing industry category dimension is |80-90| / 80 = 12.5%.

[0104] In some examples, a fifth indicator under the i-th category dimension is obtained based on the first and second quantities under the i-th category dimension, including obtaining the absolute difference between the first and second quantities under the i-th category dimension and using the absolute difference as the fifth indicator under the i-th category dimension.

[0105] In some examples, the first indicator is derived from the average of the fifth indicator across N category dimensions, including using the average of the fifth indicator across N category dimensions as the first indicator. For instance, the calculation process for the first indicator is as follows:

[0106]

[0107] Among them, C is the first indicator, C i This is the fifth indicator under the i-th category dimension.

[0108] In some examples, a first indicator is obtained based on a first quantity and a second quantity, including inputting the first quantity and the second quantity into a setting algorithm, and the setting algorithm outputting the first indicator.

[0109] In some examples, a first index is obtained based on a first quantity and a second quantity, including obtaining a first variance of multiple first quantities and a second variance of multiple second quantities, and obtaining the first index based on the first variance and the second variance.

[0110] In some examples, a first indicator is obtained based on a first quantity and a second quantity, including obtaining a first average of multiple first quantities and a second average of multiple second quantities, and obtaining the first indicator based on the first average and the second average.

[0111] S204. Based on the first evaluation result set and the second evaluation result set, the second indicator is obtained.

[0112] In the embodiments of this disclosure, the second index is obtained based on the first evaluation result set and the second evaluation result set, and may include the following possible implementation methods:

[0113] Method 1: Obtain the second indicator based on data of the same dimension from the first evaluation result set and the second evaluation result set.

[0114] It should be noted that there are no excessive restrictions on the dimensions; for example, they may include credit rating, failure rate, risk level, etc.

[0115] In some examples, a second indicator is obtained based on data of the same dimension in both the first and second evaluation result sets. This includes obtaining the second indicator by averaging multiple differences based on data of the same dimension in both the first and second evaluation result sets.

[0116] Method 2: Obtain the third number of objects under each level dimension in the first evaluation result set, and obtain the fourth number of objects under each level dimension in the second evaluation result set. Based on the third and fourth numbers, obtain the second indicator.

[0117] Therefore, this method can comprehensively consider the third number of objects under each level dimension in the first evaluation result set and the fourth number of objects under each level dimension in the second evaluation result set to obtain the second index.

[0118] It should be noted that the first and second evaluation result sets include the rating dimensions to which the object belongs. Rating dimensions may include credit rating, performance rating, risk rating, etc.

[0119] It is understandable that the number of third items may be different or the same under different level dimensions, and the number of fourth items may be different or the same under different level dimensions; no further restrictions are imposed here.

[0120] In some examples, the second indicator is obtained based on the third and fourth quantities. This includes obtaining the sixth indicator for the i-th category dimension based on the variance of the third quantity across multiple level dimensions belonging to the i-th category dimension, obtaining the seventh indicator for the i-th category dimension based on the variance of the fourth quantity across multiple level dimensions belonging to the i-th category dimension, and obtaining the second indicator based on the variances of the sixth indicator across N category dimensions and the variances of the seventh indicator across N category dimensions. Here, i is a positive integer not greater than N, and N is a positive integer. Thus, in this method, the sixth indicator for a category dimension can be obtained based on the variance of the third quantity across multiple level dimensions belonging to the same category dimension, the seventh indicator for a category dimension can be obtained based on the variance of the fourth quantity across multiple level dimensions belonging to the same category dimension, and the second indicator can be obtained based on the variances of the sixth and seventh indicators across multiple category dimensions.

[0121] It should be noted that there is a subordinate relationship between category dimensions and rank dimensions; multiple rank dimensions can belong to the same category dimension.

[0122] In some examples, a sixth indicator for the i-th category dimension is obtained based on the variance of the third number of levels belonging to the i-th category dimension, including using the variance of the third number of levels belonging to the i-th category dimension as the sixth indicator for the i-th category dimension.

[0123] For example, taking the object evaluation model as a user credit evaluation model, the category dimension is the industry category to which the user belongs, and the level dimension is the credit level to which the user belongs. If the credit level dimension belonging to the sales industry includes the first credit level, the second credit level, and the third credit level, and the third quantity under the first credit level, the second credit level, and the third credit level belonging to the sales industry are Y1, Y2, and Y3 respectively, then the sixth indicator under the sales industry category dimension is the variance of Y1, Y2, and Y3.

[0124] In some examples, the seventh index for the i-th category dimension is obtained based on the variance of the fourth number of multiple level dimensions belonging to the i-th category dimension, including using the variance of the fourth number of multiple level dimensions belonging to the i-th category dimension as the seventh index for the i-th category dimension.

[0125] For example, taking the object evaluation model as a user credit evaluation model, the category dimension is the industry category to which the user belongs, and the level dimension is the credit level to which the user belongs. If the credit level dimension belonging to the sales industry includes the first credit level, the second credit level, and the third credit level, and the fourth quantity under the first credit level, the second credit level, and the third credit level belonging to the sales industry is Y4, Y5, and Y6 respectively, then the seventh indicator under the sales industry category dimension is the variance of Y4, Y5, and Y6.

[0126] In some examples, a second indicator is obtained based on the variance of the sixth indicator under N category dimensions and the variance of the seventh indicator under N category dimensions, including the difference between the variances of the sixth indicator under N category dimensions and the variances of the seventh indicator under N category dimensions.

[0127] For example, the calculation process for the second indicator is as follows:

[0128]

[0129] Where R is the second indicator, V0 is the variance of the sixth indicator under N category dimensions, and V1 is the variance of the seventh indicator under N category dimensions.

[0130] In some cases, a second indicator is obtained based on the third and fourth quantities, including inputting the third and fourth quantities into a setting algorithm, which then outputs the second indicator.

[0131] In some examples, a second indicator is obtained based on the third and fourth quantities, including obtaining the third variance of multiple third quantities and the fourth variance of multiple fourth quantities, and obtaining the second indicator based on the third variance and the fourth variance.

[0132] In some examples, a second indicator is derived based on the third and fourth quantities, including obtaining the third average of multiple third quantities and the fourth average of multiple fourth quantities, and the second indicator is derived based on the third and fourth averages.

[0133] S205, based on the first evaluation result set, the third and fourth indicators are obtained.

[0134] In the embodiments of this disclosure, the third and fourth indicators are obtained based on the first evaluation result set, and the following possible implementation methods can be adopted:

[0135] Method 1: Obtain the third number of objects under each level dimension in the first evaluation result set. Based on the variance of the third number of objects under multiple level dimensions belonging to the i-th category dimension, obtain the sixth indicator under the i-th category dimension. Based on the skewness of the sixth indicator under N category dimensions, obtain the third indicator. Based on the kurtosis of the sixth indicator under N category dimensions, obtain the fourth indicator. Where i is a positive integer not greater than N, and N is a positive integer.

[0136] Therefore, in this method, the sixth index under the category dimension can be obtained based on the variance of the third quantity under multiple level dimensions belonging to the same category dimension, and the third index and the fourth index can be obtained based on the skewness and kurtosis of the sixth index under multiple category dimensions.

[0137] It should be noted that the first evaluation result set includes the level dimension to which the object belongs.

[0138] In some cases, a third indicator is derived based on the skewness of the sixth indicator across N category dimensions, including using the skewness of the sixth indicator across N category dimensions as the third indicator.

[0139] In some examples, the fourth indicator is derived based on the kurtosis of the sixth indicator across N category dimensions, including using the kurtosis of the sixth indicator across N category dimensions as the fourth indicator.

[0140] For example, the calculation process for the third and fourth indicators is as follows:

[0141]

[0142]

[0143] Among them, Sk is the third indicator, KT is the fourth indicator, and V 0iThis is the sixth indicator under the i-th category dimension. S is the average of the sixth indicator across N category dimensions, and S is the standard deviation of the sixth indicator across N category dimensions.

[0144] Method 2: Input the first evaluation result set into the set algorithm, and the set algorithm will output the third and fourth indicators.

[0145] Method 3: Based on the evaluation results of the same dimension in the first evaluation result set, the third and fourth indicators are obtained.

[0146] In some examples, the third and fourth indicators are obtained based on the evaluation results of the same dimension in the first evaluation result set, including the average or variance of the evaluation results of the same dimension in the first evaluation result set.

[0147] S206. Based on at least one of the first, second, third, and fourth indicators, determine whether the model update conditions are met.

[0148] In one implementation, determining whether a model update condition is met based on at least one of a first indicator, a second indicator, a third indicator, and a fourth indicator includes determining whether the first indicator, the second indicator, the third indicator, and the fourth indicator are greater than or equal to their respective set thresholds. If at least one of the first indicator, the second indicator, the third indicator, and the fourth indicator is greater than or equal to its respective set threshold, it is determined that the model update condition is met. If all of the first indicator, the second indicator, the third indicator, and the fourth indicator are less than their respective set thresholds, it is determined that the model update condition is not met.

[0149] In one implementation, determining whether the model update conditions are met based on at least one of the first, second, third, and fourth indicators includes obtaining a target indicator based on at least one of the first, second, third, and fourth indicators, and determining whether the model update conditions are met based on the target indicator.

[0150] For example, the calculation process for the target indicator is as follows:

[0151]

[0152] Where Sta is the target indicator, C is the first indicator, R is the second indicator, Sk is the third indicator, KT is the fourth indicator, and a, b, c, and d are all coefficients.

[0153] It should be noted that the relevant content on determining whether the model update conditions are met based on the target indicators can be found in the above embodiments, and will not be repeated here.

[0154] S207. If the model update conditions are met, update the object evaluation model.

[0155] The details of step S207 can be found in the above embodiments and will not be repeated here.

[0156] In summary, according to the object evaluation model update method provided in this disclosure, a first indicator is obtained based on a first dataset and a second dataset; a second indicator is obtained based on a first evaluation result set and a second evaluation result set; a third indicator and a fourth indicator are obtained based on the first evaluation result set; and a determination is made based on at least one of the first, second, third, and fourth indicators to determine whether the model update conditions are met. Therefore, the first to fourth indicators can be obtained by comprehensively considering the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set, and a determination is made based on comprehensively considering at least one of the first to fourth indicators to determine whether the model update conditions are met.

[0157] Figure 3 This is a flowchart illustrating the method for updating an object evaluation model according to a third embodiment of the present disclosure.

[0158] like Figure 3 As shown, the method for updating the object evaluation model according to the third embodiment of this disclosure includes the following steps:

[0159] S301, Obtain the first dataset and input the first dataset into the object evaluation model, and the object evaluation model outputs the first evaluation result set.

[0160] S302, obtain the second dataset after updating the first dataset, input the second dataset into the object evaluation model, and the object evaluation model outputs the second evaluation result set.

[0161] S303, based on the first evaluation result set and the second evaluation result set, determine whether the model update conditions are met.

[0162] The relevant content of steps S301-S303 can be found in the above embodiments, and will not be repeated here.

[0163] S304. If the model update condition is met, based on the first dataset and the second dataset, obtain the fifth index under N category dimensions, where N is a positive integer.

[0164] In one implementation, a fifth indicator under N category dimensions is obtained based on a first dataset and a second dataset. This includes obtaining a first number of objects under each category dimension in the first dataset, obtaining a second number of objects under each category dimension in the second dataset, and obtaining a fifth indicator under the i-th category dimension based on the first and second numbers under the i-th category dimension, where i is a positive integer not greater than N, in order to obtain the fifth indicator.

[0165] It should be noted that the relevant content of the fifth indicator can be found in the above embodiments, and will not be repeated here.

[0166] S305, based on the first evaluation result set, obtains the sixth indicator under N category dimensions.

[0167] In one implementation, a sixth indicator under N category dimensions is obtained based on a first evaluation result set. This includes obtaining the third number of objects under each level dimension in the first evaluation result set, and obtaining the sixth indicator under the i-th category dimension based on the variance of the third number under multiple level dimensions belonging to the i-th category dimension, where i is a positive integer not greater than N, in order to obtain the sixth indicator.

[0168] It should be noted that the relevant content of the sixth indicator can be found in the above embodiments, and will not be repeated here.

[0169] S306, based on the fifth indicator and the sixth indicator under N category dimensions, selects the target category dimension from the N category dimensions.

[0170] It should be noted that the number of target category dimensions is at least one.

[0171] In the embodiments of this disclosure, the target category dimension is selected from the N category dimensions based on the fifth indicator and the sixth indicator under the N category dimensions. This can include the following possible implementation methods:

[0172] Method 1: Sort the N category dimensions in descending order according to the fifth criterion to obtain the first sorting result. Then, sort the N category dimensions in descending order according to the sixth criterion to obtain the second sorting result. The top m category dimensions from the first sorting result and the top t category dimensions from the second sorting result are determined as target category dimensions. Here, m and t are both positive integers not greater than N.

[0173] Therefore, in this method, multiple category dimensions with larger fifth indicators and multiple category dimensions with larger sixth indicators can be determined as target category dimensions.

[0174] It should be noted that no restrictions are placed on m and t. For example, m = t = 0.3N. If N is 10, then m = t = 0.3 * 10 = 3.

[0175] Method 2: Based on the fifth and sixth indicators under the i-th category dimension, obtain the eighth indicator under the i-th category dimension. Based on the eighth indicator under N category dimensions, filter out the target category dimension from the N category dimensions.

[0176] In some examples, the eighth indicator in the i-th category dimension is obtained based on the fifth and sixth indicators in the i-th category dimension, including by taking a weighted average of the fifth and sixth indicators in the i-th category dimension.

[0177] In some examples, the target category dimension is selected from the N category dimensions based on the eighth indicator under the N category dimensions. This includes sorting the N category dimensions from largest to smallest according to the eighth indicator to obtain a third sorting result, and determining the top p category dimensions in the third sorting result as the target category dimension.

[0178] S307, determine the first model parameter associated with the target category dimension from the model parameters of the object evaluation model.

[0179] It's important to note that there's a correlation between the model parameters of the object rating model and the category dimension. The rating dimension to which an object belongs under a category dimension is derived based on the model parameters of the object rating model associated with that category dimension. For example, taking a user credit rating model as an example, the category dimension represents the user's industry category, and the rating dimension represents the user's credit rating. The credit rating of a user under the sales industry category dimension is derived based on the model parameters of the object rating model associated with that sales industry category dimension, and the credit rating of a user under the manufacturing industry category dimension is derived based on the model parameters of the object rating model associated with that manufacturing industry category dimension.

[0180] It is understandable that different category dimensions can be associated with different model parameters or the same model parameters; we will not impose further restrictions here. The first model parameter refers to the model parameters in the object evaluation model that are associated with the target category dimension.

[0181] In one implementation, a first model parameter associated with the target category dimension is determined from the model parameters of the object evaluation model. This includes determining the first model parameter associated with the target category dimension based on the correlation between the model parameters of the object evaluation model and the category dimension. It is understood that the aforementioned correlation can be preset.

[0182] In one implementation, the object evaluation model may include sub-models that correspond one-to-one with the category dimensions. From the model parameters of the object evaluation model, a first model parameter associated with the target category dimension is determined, including determining the model parameters of the sub-models corresponding to the target category dimension as the first model parameter.

[0183] S308, Update the parameters of the first model.

[0184] In one implementation, updating the first model parameters includes obtaining training samples, wherein the training samples include sample data and sample evaluation results under the target category dimension, and updating the first model parameters based on the training samples.

[0185] In one implementation, updating the first model parameters includes obtaining model parameters of at least one set model, and performing a weighted summation of the first model parameters and the model parameters of the at least one set model to obtain the updated second model parameters. Therefore, the first model parameters can be updated by comprehensively considering the first model parameters and the model parameters of the at least one set model, improving the accuracy and efficiency of model parameter updates.

[0186] It should be noted that the weights of the first model parameters and the model parameters of at least one set model are not subject to too much restriction. For example, if the set model includes a DNN model and an LR model, the weights of the first model parameters and the model parameters of the DNN model are greater than or equal to the model parameters of the LR model.

[0187] In summary, according to the object evaluation model update method provided in this disclosure, a fifth indicator under N category dimensions is obtained based on the first dataset and the second dataset; a sixth indicator under N category dimensions is obtained based on the first evaluation result set; a target category dimension is selected from the N category dimensions based on the fifth and sixth indicators; and a first model parameter associated with the target category dimension is determined from the model parameters of the object evaluation model, and the first model parameter is updated. Therefore, the target category dimension can be selected from the N category dimensions based on the fifth and sixth indicators, and only the first model parameter associated with the target category dimension needs to be updated, i.e., only a portion of the model parameters of the object evaluation model needs to be updated, thus improving the update accuracy and efficiency of the object evaluation model.

[0188] Based on any of the above embodiments, such as Figure 4 As shown, a first dataset can be obtained and input into the object evaluation model, which outputs a first evaluation result set. Then, a second dataset updated from the first dataset can be obtained and input into the object evaluation model, which outputs a second evaluation result set.

[0189] Based on the first dataset and the second dataset, a first indicator is obtained. Based on the first evaluation result set and the second evaluation result set, a second indicator is obtained. Based on the first evaluation result set, a third indicator and a fourth indicator are obtained. Based on the first indicator, the second indicator, the third indicator and the fourth indicator, a target indicator Sta is obtained. If the target indicator Sta is greater than or equal to a set threshold, it is determined that the model update condition is not met. If the target indicator Sta is less than the set threshold, it is determined that the model update condition is met. Based on the model parameters of at least one set model, the object evaluation model is updated.

[0190] It should be noted that the relevant content on updating the object evaluation model based on at least one set model parameter can be found in the above embodiments, and will not be repeated here.

[0191] Figure 5 This is a flowchart illustrating the method for updating an object evaluation model according to the first embodiment of this disclosure.

[0192] like Figure 5 As shown, the object evaluation model update method of the first embodiment of this disclosure includes the following steps:

[0193] S501, retrieve the target data of the object.

[0194] S502, the target data is input into the object evaluation model, and the object evaluation model outputs the object evaluation result. The object evaluation model is obtained by using the object evaluation model update method.

[0195] It should be noted that the object evaluation model update method of this disclosure is executed by an electronic device. The object evaluation model update method of this disclosure can be executed by the object evaluation model update device of this disclosure, which can be configured in any electronic device to execute the object evaluation model update method of this disclosure.

[0196] It should be noted that the relevant content of the target data can be referred to the relevant content of the first data in the above embodiments, and the relevant content of the object evaluation results and object evaluation model can be referred to the above embodiments, which will not be repeated here.

[0197] It should be noted that the object evaluation model can adopt... Figures 1 to 4 The update method for the object evaluation model is obtained.

[0198] For example, taking a user credit rating model as an example, user feature data can be input into the user credit rating model, and the user credit rating model can output the user's credit rating, credit score, etc. That is, the target data can include user feature data, and the evaluation results of the object can include the user's credit rating, credit score, etc.

[0199] For example, taking a product performance evaluation model as an example, the product's feature data can be input into the product performance evaluation model, and the product performance evaluation model can output the product's performance level, performance score, failure rate, etc. That is, the target data can include the product's feature data, and the evaluation results can include the product's performance level, performance score, failure rate, etc.

[0200] For example, taking an object evaluation model as a project risk assessment model, the project's characteristic data can be input into the project's risk assessment model, and the model can output the project's risk level, risk score, etc. In other words, the target data can include the project's characteristic data, and the object evaluation results can include the project's risk level, risk score, etc.

[0201] In summary, according to the object evaluation method provided in this embodiment, target data of the object is obtained, the target data is input into the object evaluation model, and the object evaluation model outputs the evaluation result of the object. The object evaluation model is obtained by using an object evaluation model update method, which can realize continuous updating of the model during the use of the object evaluation model. The object evaluation model has high accuracy, and thus the object evaluation has high accuracy, which is suitable for user credit evaluation scenarios.

[0202] Figure 6 This is a block diagram of an apparatus for updating an object evaluation model according to a first embodiment of the present disclosure.

[0203] like Figure 6 As shown, the object evaluation model update device 600 of this embodiment includes: a first acquisition module 601, a second acquisition module 602, a judgment module 603 and an update module 604.

[0204] The first acquisition module 601 is configured to acquire a first dataset, input the first dataset into the object evaluation model, and output a first evaluation result set from the object evaluation model.

[0205] The second acquisition module 602 is configured to acquire the second dataset after the first dataset is updated, input the second dataset into the object evaluation model, and output the second evaluation result set by the object evaluation model;

[0206] The judgment module 603 is configured to determine whether the model update conditions are met based on the first evaluation result set and the second evaluation result set.

[0207] The update module 604 is configured to update the object evaluation model if the model update conditions are met.

[0208] In one embodiment of this disclosure, the judgment module 603 is further configured to: determine whether the model update condition is met based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set.

[0209] In one embodiment of this disclosure, the judgment module 603 is further configured to: obtain a first indicator based on the first dataset and the second dataset; obtain a second indicator based on the first evaluation result set and the second evaluation result set; obtain a third indicator and a fourth indicator based on the first evaluation result set; and determine whether the model update condition is met based on at least one of the first indicator, the second indicator, the third indicator, and the fourth indicator.

[0210] In one embodiment of this disclosure, the first dataset and the second dataset include category dimensions to which the objects belong, and the judgment module 603 is further configured to: obtain a first number of objects under each category dimension in the first dataset; obtain a second number of objects under each category dimension in the second dataset; and obtain the first index based on the first number and the second number.

[0211] In one embodiment of this disclosure, the judgment module 603 is further configured to: obtain a fifth indicator under the i-th category dimension based on the first quantity and the second quantity under the i-th category dimension; and obtain the first indicator based on the average of the fifth indicators under N category dimensions; wherein i is a positive integer not greater than N, and N is a positive integer.

[0212] In one embodiment of this disclosure, the first evaluation result set and the second evaluation result set include the level dimension to which the object belongs. The judgment module 603 is further configured to: obtain a third number of objects under each level dimension in the first evaluation result set; obtain a fourth number of objects under each level dimension in the second evaluation result set; and obtain the second indicator based on the third number and the fourth number.

[0213] In one embodiment of this disclosure, the judgment module 603 is further configured to: obtain a sixth index under the i-th category dimension based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension; obtain a seventh index under the i-th category dimension based on the variance of the fourth quantity under multiple level dimensions belonging to the i-th category dimension; and obtain a second index based on the variance of the sixth index under N category dimensions and the variance of the seventh index under N category dimensions; wherein i is a positive integer not greater than N, and N is a positive integer.

[0214] In one embodiment of this disclosure, the first evaluation result set includes the level dimension to which the object belongs, and the judgment module 603 is further configured to: obtain the third quantity of objects under each level dimension in the first evaluation result set; obtain the sixth index under the i-th category dimension based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension; obtain the third index based on the skewness of the sixth index under N category dimensions; and obtain the fourth index based on the kurtosis of the sixth index under N category dimensions; wherein i is a positive integer not greater than N, and N is a positive integer.

[0215] In one embodiment of this disclosure, the first dataset and the second dataset include the category dimension to which the object belongs, and the first evaluation result set includes the level dimension to which the object belongs; wherein, the update module 604 is further configured to: obtain a fifth indicator under N category dimensions based on the first dataset and the second dataset, where N is a positive integer; obtain a sixth indicator under N category dimensions based on the first evaluation result set; filter out a target category dimension from the N category dimensions based on the fifth indicator and the sixth indicator under N category dimensions; determine a first model parameter associated with the target category dimension from the model parameters of the object evaluation model; and update the first model parameter.

[0216] In one embodiment of this disclosure, the update module 604 is further configured to: obtain a first number of objects under each category dimension in the first dataset; obtain a second number of objects under each category dimension in the second dataset; and obtain a fifth index under the i-th category dimension based on the first number and the second number under the i-th category dimension, where i is a positive integer not greater than N.

[0217] In one embodiment of this disclosure, the update module 604 is further configured to: obtain the third number of objects under each level dimension in the first evaluation result set; and obtain the sixth index under the i-th category dimension based on the variance of the third number under multiple level dimensions belonging to the i-th category dimension, where i is a positive integer not greater than N.

[0218] In one embodiment of this disclosure, the update module 604 is further configured to: obtain model parameters of at least one set model; and perform a weighted summation of the first model parameters and the model parameters of at least one set model to obtain the second model parameters after the first model parameters are updated.

[0219] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0220] In summary, the object evaluation model update apparatus provided in this embodiment acquires a first dataset and inputs it into the object evaluation model, which outputs a first evaluation result set. It then acquires a second dataset updated from the first dataset, inputs it into the object evaluation model, and outputs a second evaluation result set. Based on the first and second evaluation result sets, it determines whether the model update conditions are met. If the conditions are met, the object evaluation model is updated. Therefore, by comprehensively considering both the first and second evaluation result sets, the apparatus automatically determines whether the model update conditions are met and updates the object evaluation model when they are met. This achieves automatic updating of the object evaluation model, enabling continuous model updates during its use and improving its accuracy. This apparatus is suitable for updating user credit evaluation models.

[0221] Figure 7 This is a block diagram of an object evaluation apparatus according to a first embodiment of the present disclosure.

[0222] like Figure 7 As shown, the object evaluation device 700 of this embodiment includes: an acquisition module 701 and an evaluation module 702.

[0223] Module 701 is configured to retrieve the target data of an object.

[0224] Evaluation module 702 is configured to input the target data into an object evaluation model, and output the evaluation result of the object by the object evaluation model, wherein the object evaluation model is obtained by the object evaluation model update method as described in the first aspect above.

[0225] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0226] In summary, the object evaluation device provided in this embodiment acquires target data of an object, inputs the target data into an object evaluation model, and outputs the evaluation result of the object by the object evaluation model. The object evaluation model is obtained by an object evaluation model update method, which can realize continuous updating of the model during the use of the object evaluation model. The object evaluation model has high accuracy, and thus the object evaluation has high accuracy, which is suitable for user credit evaluation scenarios.

[0227] Figure 8 This is a block diagram illustrating an electronic device according to an exemplary embodiment.

[0228] like Figure 8As shown, the above-mentioned electronic device 800 includes:

[0229] The memory 810 and processor 820 are connected by a bus 830, which connects different components (including the memory 810 and the processor 820). The memory 810 stores a computer program, which, when executed by the processor 820, implements the object evaluation model update method and / or object evaluation method described in the embodiments of this disclosure.

[0230] Bus 830 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. Examples of these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0231] Electronic device 800 typically includes a variety of electronic device readable media. These media can be any available media that can be accessed by electronic device 800, including volatile and non-volatile media, removable and non-removable media.

[0232] The memory 810 may also include computer system readable media in the form of volatile memory, such as random access memory (RAM) 840 and / or cache memory 850. The electronic device 800 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 860 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 8 Not shown; usually referred to as a "hard drive"). Although Figure 8 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 830 via one or more data media interfaces. Memory 810 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of this disclosure.

[0233] A program / utility 880 having a set (at least one) of program modules 870 may be stored, for example, in memory 810. Such program modules 870 include—but are not limited to—an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 870 typically perform the functions and / or methods described in the embodiments of this disclosure.

[0234] Electronic device 800 can also communicate with one or more external devices 890 (e.g., keyboard, pointing device, display 891, etc.), and with one or more devices that enable a user to interact with the electronic device 800, and / or with any device that enables the electronic device 800 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed through input / output (I / O) interface 892. Furthermore, electronic device 800 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) through network adapter 893. Figure 8 As shown, network adapter 893 communicates with other modules of electronic device 800 via bus 830. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with electronic device 800, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0235] The processor 820 performs various functional applications and data processing by running programs stored in the memory 810.

[0236] It should be noted that the implementation process and technical principles of the electronic device in this embodiment are explained in the foregoing description of the object evaluation model update method and object evaluation method in the embodiments of this disclosure, and will not be repeated here.

[0237] In summary, the electronic device provided in this embodiment can execute the object evaluation model update method and / or object evaluation method as described above, obtain a first dataset, input the first dataset into the object evaluation model, and have the object evaluation model output a first evaluation result set. It then obtains a second dataset updated from the first dataset, inputs the second dataset into the object evaluation model, and has the object evaluation model output a second evaluation result set. Based on the first and second evaluation result sets, it determines whether the model update conditions are met. If the model update conditions are met, the object evaluation model is updated. Therefore, by comprehensively considering both the first and second evaluation result sets, it can determine whether the model update conditions are met, thus achieving automatic determination of the model update conditions. When the model update conditions are met, the object evaluation model is updated, achieving automatic updating of the object evaluation model. This allows for continuous updating of the model during its use, improving the accuracy of the object evaluation model and making it suitable for updating user credit evaluation models.

[0238] To implement the above embodiments, this disclosure also proposes a computer-readable storage medium.

[0239] When the instructions in the computer-readable storage medium are executed by the processor of an electronic device, the electronic device is able to perform the object evaluation model update method and / or object evaluation method as described above. Optionally, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0240] To implement the above embodiments, this disclosure also provides a computer program product, including a computer program, characterized in that, when the computer program is executed by a processor, it implements the object evaluation model update method and / or object evaluation method as described above.

[0241] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0242] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for updating an object evaluation model, characterized in that, The object evaluation model is a product performance evaluation model, and the method includes: A first dataset is obtained and input into the object evaluation model, and the object evaluation model outputs a first evaluation result set, wherein the first dataset includes the feature data of the product and the first evaluation result set includes the performance evaluation result data of the product. Obtain the second dataset updated from the first dataset, and input the second dataset into the object evaluation model. The object evaluation model then outputs a second evaluation result set. The first dataset and the second dataset include the category dimension to which the object belongs, and the first evaluation result set and the second evaluation result set include the level dimension to which the object belongs. Based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set, determine whether the model update conditions are met, including: The first metric is obtained based on the number of objects in each category dimension in the first dataset and the second dataset; Based on the number of objects in each level dimension of the first evaluation result set and the second evaluation result set, a second indicator is obtained; Based on the number of objects in each level dimension of the first evaluation result set, a third indicator and a fourth indicator are obtained. The third indicator represents the skewness of the variance distribution of the object quantity distribution, and the fourth indicator represents the kurtosis of the variance distribution of the object quantity distribution. Based on at least one of the first indicator, the second indicator, the third indicator, and the fourth indicator, determine whether the model update condition is met; If the model update conditions are met, the object evaluation model is updated.

2. The method according to claim 1, characterized in that, Based on the first dataset and the second dataset, a first metric is obtained, including: Obtain the first number of objects in each category dimension of the first dataset; Obtain the second number of objects in each category dimension of the second dataset; The first indicator is obtained based on the first quantity and the second quantity.

3. The method according to claim 2, characterized in that, The process of obtaining the first indicator based on the first quantity and the second quantity includes: Based on the first quantity and the second quantity under the i-th category dimension, the fifth indicator under the i-th category dimension is obtained; The first indicator is obtained based on the average of the fifth indicator across N category dimensions; where, i is a positive integer not greater than N, and N is a positive integer.

4. The method according to claim 1, characterized in that, Based on the first evaluation result set and the second evaluation result set, a second indicator is obtained, including: Obtain the third number of objects under each level dimension in the first evaluation result set; Obtain the fourth number of objects under each level dimension in the second evaluation result set; The second indicator is obtained based on the third and fourth quantities.

5. The method according to claim 4, characterized in that, The process of obtaining the second indicator based on the third quantity and the fourth quantity includes: Based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension, the sixth index under the i-th category dimension is obtained; Based on the variance of the fourth quantity under multiple level dimensions belonging to the i-th category dimension, the seventh index under the i-th category dimension is obtained; The second indicator is obtained based on the variance of the sixth indicator across N category dimensions and the variance of the seventh indicator across N category dimensions; where, i is a positive integer not greater than N, and N is a positive integer.

6. The method according to claim 1, characterized in that, Based on the first evaluation result set, the third and fourth indicators are obtained, including: Obtain the third number of objects under each level dimension in the first evaluation result set; Based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension, the sixth index under the i-th category dimension is obtained; The third indicator is obtained based on the skewness of the sixth indicator under N category dimensions; The fourth indicator is obtained based on the kurtosis of the sixth indicator across N category dimensions; where, i is a positive integer not greater than N, and N is a positive integer.

7. The method according to any one of claims 1-6, characterized in that, The updating of the object evaluation model includes: Based on the first dataset and the second dataset, a fifth indicator is obtained under N category dimensions, where N is a positive integer; Based on the first evaluation result set, a sixth indicator under N category dimensions is obtained; Based on the fifth indicator and the sixth indicator under N category dimensions, the target category dimension is selected from the N category dimensions; From the model parameters of the object evaluation model, determine the first model parameter associated with the target category dimension; Update the parameters of the first model.

8. The method according to claim 7, characterized in that, The fifth indicator, based on the first dataset and the second dataset, is obtained under N category dimensions, including: Obtain the first number of objects in each category dimension of the first dataset; Obtain the second number of objects in each category dimension of the second dataset; Based on the first quantity and the second quantity under the i-th category dimension, the fifth index under the i-th category dimension is obtained, where i is a positive integer not greater than N.

9. The method according to claim 7, characterized in that, Based on the first evaluation result set, the sixth indicator under N category dimensions is obtained, including: Obtain the third number of objects under each level dimension in the first evaluation result set; Based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension, the sixth index under the i-th category dimension is obtained, where i is a positive integer not greater than N.

10. The method according to claim 7, characterized in that, The step of updating the first model parameters includes: Obtain at least one model parameter of the specified model; The first model parameters and the model parameters of at least one of the set models are weighted and summed to obtain the second model parameters after the first model parameters are updated.

11. An object evaluation method, characterized in that, include: Retrieve the target data of the object; The target data is input into the object evaluation model, and the object evaluation model outputs the evaluation result of the object. The object evaluation model is obtained by the object evaluation model update method as described in any one of claims 1-10.

12. An update device for an object evaluation model, characterized in that, The object evaluation model is a product performance evaluation model, and the device includes: The first acquisition module is configured to acquire a first dataset and input the first dataset into the object evaluation model, and output a first evaluation result set by the object evaluation model, wherein the first dataset includes the feature data of the product and the first evaluation result set includes the performance evaluation result data of the product. The second acquisition module is configured to acquire the second dataset after the first dataset is updated, and input the second dataset into the object evaluation model, and output the second evaluation result set by the object evaluation model. The first dataset and the second dataset include the category dimension to which the object belongs, and the first evaluation result set and the second evaluation result set include the level dimension to which the object belongs. The judgment module is configured to determine whether the model update conditions are met based on the first dataset, the second dataset, the first evaluation result set, and the second evaluation result set. The update module is configured to update the object evaluation model if the model update conditions are met. The judgment module is further configured as follows: The first metric is obtained based on the number of objects in each category dimension in the first dataset and the second dataset; Based on the number of objects in each level dimension of the first evaluation result set and the second evaluation result set, a second indicator is obtained; Based on the number of objects in each level dimension of the first evaluation result set, a third indicator and a fourth indicator are obtained. The third indicator represents the skewness of the variance distribution of the object quantity distribution, and the fourth indicator represents the kurtosis of the variance distribution of the object quantity distribution. Based on at least one of the first indicator, the second indicator, the third indicator, and the fourth indicator, determine whether the model update condition is met.

13. The apparatus according to claim 12, characterized in that, The judgment module is further configured to: Obtain the first number of objects in each category dimension of the first dataset; Obtain the second number of objects in each category dimension of the second dataset; The first indicator is obtained based on the first quantity and the second quantity.

14. The apparatus according to claim 13, characterized in that, The judgment module is further configured to: Based on the first quantity and the second quantity under the i-th category dimension, the fifth indicator under the i-th category dimension is obtained; The first indicator is obtained based on the average of the fifth indicator across N category dimensions; where, i is a positive integer not greater than N, and N is a positive integer.

15. The apparatus according to claim 12, characterized in that, The judgment module is further configured to: Obtain the third number of objects under each level dimension in the first evaluation result set; Obtain the fourth number of objects under each level dimension in the second evaluation result set; The second indicator is obtained based on the third and fourth quantities.

16. The apparatus according to claim 15, characterized in that, The judgment module is further configured to: Based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension, the sixth index under the i-th category dimension is obtained; Based on the variance of the fourth quantity under multiple level dimensions belonging to the i-th category dimension, the seventh index under the i-th category dimension is obtained; The second indicator is obtained based on the variance of the sixth indicator across N category dimensions and the variance of the seventh indicator across N category dimensions; where, i is a positive integer not greater than N, and N is a positive integer.

17. The apparatus according to claim 12, characterized in that, The judgment module is further configured to: Obtain the third number of objects under each level dimension in the first evaluation result set; Based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension, the sixth index under the i-th category dimension is obtained; The third indicator is obtained based on the skewness of the sixth indicator under N category dimensions; The fourth indicator is obtained based on the kurtosis of the sixth indicator across N category dimensions; where, i is a positive integer not greater than N, and N is a positive integer.

18. The apparatus according to any one of claims 12-17, characterized in that, The update module is also configured to: Based on the first dataset and the second dataset, a fifth indicator is obtained under N category dimensions, where N is a positive integer; Based on the first evaluation result set, a sixth indicator under N category dimensions is obtained; Based on the fifth indicator and the sixth indicator under N category dimensions, the target category dimension is selected from the N category dimensions; From the model parameters of the object evaluation model, determine the first model parameter associated with the target category dimension; Update the parameters of the first model.

19. The apparatus according to claim 18, characterized in that, The update module is also configured to: Obtain the first number of objects in each category dimension of the first dataset; Obtain the second number of objects in each category dimension of the second dataset; Based on the first quantity and the second quantity under the i-th category dimension, the fifth index under the i-th category dimension is obtained, where i is a positive integer not greater than N.

20. The apparatus according to claim 18, characterized in that, The update module is also configured to: Obtain the third number of objects under each level dimension in the first evaluation result set; Based on the variance of the third quantity under multiple level dimensions belonging to the i-th category dimension, the sixth index under the i-th category dimension is obtained, where i is a positive integer not greater than N.

21. The apparatus according to claim 18, characterized in that, The update module is also configured to: Obtain at least one model parameter of the specified model; The first model parameters and the model parameters of at least one of the set models are weighted and summed to obtain the second model parameters after the first model parameters are updated.

22. An object evaluation device, characterized in that, include: The retrieval module is configured to retrieve the target data of an object; The evaluation module is configured to input the target data into an object evaluation model, and the object evaluation model outputs the evaluation result of the object, wherein the object evaluation model is obtained using the object evaluation model update method as described in any one of claims 1-10.

23. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the method as described in any one of claims 1-11.

24. A computer-readable storage medium, wherein instructions in the computer-readable storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the method as described in any one of claims 1-11.

25. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-11.

Images