CN115860505A - Object evaluation method and device, terminal equipment and storage medium - Google Patents
Object evaluation method and device, terminal equipment and storage medium Download PDFInfo
- Publication number
- CN115860505A CN115860505A CN202211019789.6A CN202211019789A CN115860505A CN 115860505 A CN115860505 A CN 115860505A CN 202211019789 A CN202211019789 A CN 202211019789A CN 115860505 A CN115860505 A CN 115860505A
- Authority
- CN
- China
- Prior art keywords
- target
- evaluated
- score
- value
- preset
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application is suitable for the technical field of data processing, and provides an object evaluation method, an object evaluation device, terminal equipment and a storage medium, wherein the method comprises the following steps: acquiring first target data of the object to be evaluated in a target historical time period according to a preset evaluation type of the object to be evaluated, and determining a first grading change value of the object to be evaluated according to the first target data; acquiring target events of all target objects in a target historical time period, and determining a second grading change value of the object to be evaluated according to the target events of all the target objects; the preset evaluation type of the target object is the same as that of the object to be evaluated; and determining the target score value of the object to be evaluated in the target historical time period according to the first score change value and the second score change value. By adopting the method, the target score value with higher accuracy of the object to be evaluated can be generated.
Description
Technical Field
The present application belongs to the technical field of data processing, and in particular, to an object evaluation method, an object evaluation device, a terminal device, and a storage medium.
Background
An external evaluation of a company may typically reflect the overall operation of the company or the development of the company in a certain area. In recent years, the evaluation system and standard of the company are more and more regulated, and the company is more concerned about the evaluation of the company by a third party organization and a national organization.
Currently, an evaluation mode adopted by a third-party organization for evaluating a company is generally to acquire data under multiple evaluation indexes of the company, and perform weighted summation on the data under the multiple evaluation indexes based on weights corresponding to the evaluation indexes, so as to obtain a score value of the company, wherein the score value can reflect the quality of external evaluation of the company. However, the weight corresponding to each evaluation index is usually determined by a plurality of experts, and for the same evaluation index, the weights given by different experts are often different, so that the evaluation standards of companies are not uniform, and the weight of the evaluation index given by an expert is greatly influenced by subjective factors. Therefore, when the evaluation is carried out only according to the data of a plurality of indexes of the company, the finally calculated evaluation value of the company is not objective and accurate enough, and the evaluation of the company is relatively smooth.
Disclosure of Invention
The embodiment of the application provides an object evaluation method, an object evaluation device, terminal equipment and a storage medium, and can solve the problem that evaluation is only carried out according to data of multiple indexes of a company, and the obtained score value of the company is not objective and accurate enough.
In a first aspect, an embodiment of the present application provides an object evaluation method, including:
according to the preset evaluation type of the object to be evaluated, first target data of the object to be evaluated in a target historical time period are obtained, and a first score change value of the object to be evaluated is determined according to the first target data;
acquiring target events of all target objects in a target historical time period, and determining a second grading change value of the object to be evaluated according to the target events of all the target objects; the preset evaluation type of the target object is the same as that of the object to be evaluated;
and determining the target score value of the object to be evaluated in the target historical time period according to the first score change value and the second score change value.
In a second aspect, an embodiment of the present application provides an object evaluation apparatus, including:
the first grading change value determining module is used for acquiring first target data of the object to be evaluated in a target historical time period according to the preset evaluation type of the object to be evaluated and determining a first grading change value of the object to be evaluated according to the first target data;
the second score value determining module is used for acquiring target events of all target objects in a target historical time period and determining second score change values of the objects to be evaluated according to the target events of all the target objects; the preset evaluation type of the target object is the same as that of the object to be evaluated;
and the target score value determining module is used for determining the target score value of the object to be evaluated in the target historical time period according to the first score change value and the second score change value.
In a third aspect, an embodiment of the present application provides a terminal device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the method according to the first aspect is implemented.
In a fourth aspect, embodiments of the present application provide a computer-readable storage medium, in which a computer program is stored, and the computer program, when executed by a processor, implements the method of the first aspect.
In a fifth aspect, embodiments of the present application provide a computer program product, which, when run on a terminal device, causes the terminal device to execute the method of the first aspect.
Compared with the prior art, the embodiment of the application has the advantages that: according to the preset evaluation type of the object to be evaluated, first target data of the object to be evaluated in the target historical time period and target events of other target objects belonging to the same evaluation type in the target historical time period are respectively acquired. Then, the terminal device may determine a first score change value of the object to be evaluated according to the first target data. And meanwhile, determining a second grading change value of the object to be evaluated in the target historical time period according to the target event. Finally, the terminal device may generate a target score value of the object to be evaluated in the target window period based on the first score change value and the second score change value at the same time. Therefore, when the target scoring value is generated, not only the first target data of the object to be evaluated is only based, but also the influence possibly generated by each target event of the target objects with the same evaluation type in the target historical time period is considered. Furthermore, the target scoring value of the company can be obtained from multiple aspects, so that the finally obtained target scoring value is more objective and accurate.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a flowchart illustrating an implementation of an object evaluation method according to an embodiment of the present application;
fig. 2 is a schematic diagram of an implementation manner of obtaining first target data in an object evaluation method according to an embodiment of the present application;
fig. 3 is a schematic diagram illustrating an implementation manner of generating a scoring model in an object evaluation method according to an embodiment of the present application;
fig. 4 is a schematic diagram illustrating an implementation manner of determining a second score change value in an object evaluation method according to an embodiment of the present application;
fig. 5 is a schematic diagram of an implementation manner of obtaining a target event in an object evaluation method according to an embodiment of the present application;
fig. 6 is a schematic structural diagram of an object evaluation apparatus according to an embodiment of the present application;
fig. 7 is a schematic structural diagram of a terminal device according to an embodiment of the present application.
Detailed Description
In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
Furthermore, in the description of the present application and the appended claims, the terms "first," "second," "third," and the like are used for distinguishing between descriptions and not necessarily for describing or implying relative importance.
In recent years, as the evaluation system and standards of companies are more and more regulated, companies pay more attention to the evaluation of the companies by third parties and national institutions. Meanwhile, with the complexity of a financial system and the continuous development and perfection of an evaluation system along with the continuous change of economic environment, a single index is difficult to comprehensively evaluate the performance of a company in a certain field.
Currently, when a third-party organization evaluates a company, most of the evaluation is carried out by using an analytic hierarchy process. The method comprises the steps of acquiring data under multiple evaluation indexes of a company, and carrying out weighted summation on the data under the multiple evaluation indexes based on the weight corresponding to each evaluation index, so as to obtain a score value capable of reflecting the quality of external evaluation of the company. However, in this method, the weight of each evaluation index is generally determined by a plurality of experts. Therefore, it is greatly influenced by subjective factors. Therefore, when the evaluation is carried out only according to the data of a plurality of indexes of the company, the finally calculated evaluation value of the company is not objective and accurate enough, and the evaluation of the company is relatively smooth.
Therefore, in order to obtain an objective and accurate evaluation of a company, the present embodiment provides an object evaluation method, which may be applied to a terminal device such as a tablet computer, a notebook computer, an ultra-mobile personal computer (UMPC), a netbook, and the like, and the specific type of the terminal device is not limited in the embodiments of the present application.
In an embodiment, the object to be evaluated may be an object such as an object, an individual, or a company, and the like, which is not limited thereto. In this embodiment, the object to be evaluated may be a company. When a company is evaluated, the operation risk, the management level, the Social responsibility and the performance of the company and/or a plurality of preset evaluation types such as environment, society and Governance (ESG) and the like can be evaluated, which is not limited.
It should be noted that, when the preset evaluation types of companies in different aspects are evaluated, the data required to be used are generally different from each other. For example, when the operational risk of a company is evaluated, the data used for evaluation is generally various data generated by the company in the operational process. Such as liability rate or assets. For the evaluation of a company in the environment, the data used for the evaluation is generally various data generated by the company in the process of treating the environment. For example, the type of data such as the amount of emission pollution or the degree of air pollution is not limited.
Referring to fig. 1, fig. 1 shows a flowchart of an implementation of an object evaluation method provided in an embodiment of the present application, where the method includes the following steps:
s101, according to a preset evaluation type of an object to be evaluated, first target data of the object to be evaluated in a target historical time period are obtained, and a first grading change value of the object to be evaluated is determined according to the first target data.
In the application, the preset evaluation type is explained above, and the explanation is not provided. It should be noted that, when determining the preset evaluation type of the object to be evaluated, the third-party evaluation company may determine the preset evaluation type in the terminal device, or the terminal device may determine the preset evaluation type according to the field to which the object to be evaluated belongs, which is not limited to this.
For example, when the field to which the object to be evaluated belongs is a financial field, the preset evaluation type for evaluating the company may be considered as an operational risk.
In application, when the first target data is obtained, the terminal device may receive the first target data input by the third party company, or may query the first target data related to the object to be evaluated on the internet through a networking function, which is not limited herein.
In an application, the target historical time period may be a plurality of time periods before the current time, and the time periods have continuity. In this embodiment, the target history period may be a period in days. That is, the target history period is usually a plurality of dates before the current date, and the plurality of dates have continuity. For example, each date within one month prior to the current date may be considered a target historical time period.
In an application, the first target data is data generated by the object to be evaluated in a target history time period, and the data generated by the object to be evaluated in the target history time period is various. In this embodiment, of the plurality of types of data, data related to a preset evaluation type may be used as the first target data. Specifically, the description of the data used when the evaluation is performed with reference to the preset evaluation type of the operation risk of the company can be referred to.
It should be noted that, when data which is generated by the object to be evaluated in the target historical time period and can be used for evaluating the preset evaluation type of the object to be evaluated is acquired, the acquired data may have a deficiency. Therefore, the acquired data is also required to be filled in to generate the required first target data,
specifically, referring to fig. 2, the terminal device may specifically perform the padding on the acquired data through S201-S206 shown in fig. 2, which is detailed as follows:
s201, acquiring a plurality of initial data corresponding to a plurality of evaluation indexes of an object to be evaluated in a target historical time period.
In application, the initial data is data which is generated by the object to be evaluated in the target historical time period and can be used for evaluating the preset evaluation type of the object to be evaluated. Wherein, for the initial data, it may be due to the development condition of the company, and no data is generated in a certain period of time, or the generated data is lost when being acquired. Based on this, the terminal device needs to fill in a plurality of initial data.
And S202, if the plurality of initial data corresponding to the first evaluation index in the plurality of evaluation indexes are missing, filling the missing data in the plurality of initial data corresponding to the first evaluation index with a target value.
In application, the first evaluation index may be: the object to be evaluated generates data without relevant statistics. For example, when the preset evaluation type is the operation risk of a company, the first evaluation index may be data that has no relevant statistics actually, such as "no-flow negative value". Based on the first evaluation index data, the terminal device may fill in the missing data with the target value. In this embodiment, the target value may be 0.
In an application, the first evaluation index may be determined in advance by a third party company, and the number of the first evaluation indexes may be one or more, which is not limited.
And S203, if a plurality of initial data corresponding to a second evaluation index in the plurality of evaluation indexes are missing, determining the target date of the missing data in the plurality of initial data corresponding to the second evaluation index.
In application, the second evaluation index is different from the first evaluation index, and the initial data corresponding to the second index is usually not the data having no relevant statistics actually. Therefore, the target value cannot be generally used for filling.
In the application, the target date is a target date of missing data in the plurality of pieces of initial data corresponding to the second evaluation index. It has been described above that the target history period is generally in units of days, and therefore, when a plurality of pieces of initial data corresponding to the object to be evaluated in the target history period are acquired, the date of each piece of initial data may also be acquired at the same time. The target date is a date with initial data missing in a plurality of initial data corresponding to the second index.
And S204, if target initial data with the date adjacent to the target date exists in the plurality of initial data corresponding to the second evaluation index, filling missing data in the plurality of initial data corresponding to the second evaluation index by adopting the target initial data.
In application, the date of the target initial data is adjacent to the target date. When padding with target initial data, the target initial data generally refers to the initial data missing in the adjacent target date.
When the missing data is filled, the missing data is usually filled in order according to the target date of the missing initial data. Furthermore, the initial data on the target date after the padding cannot be used to pad the initial data missing on other target dates.
For example, for initial data at multiple dates such as dates 1, 2, 3, T, etc., if the initial data of number 2 and number 3 have a deletion, the initial data of number 2 should be filled in first. At this time, no. 2 is the target date, and No. 3 is missing. Therefore, the target initial data of the date adjacent to the target date should be the initial data corresponding to No. 1. That is, the initial data corresponding to number 1 should be filled in the initial data missing in number 2. However, when the missing initial data No. 3 is filled, the corresponding initial data No. 2 is also missing before that. If the initial data of number 2 after padding is padded in the initial data of number 3 missing, the authenticity of the padded initial data in this manner is low. Therefore, in the present embodiment, the target initial data whose date is adjacent to the target date is present in the plurality of initial data corresponding to the second evaluation index, and the initial data missing on the target date is filled.
It should be added that, when the initial data missing on the target date is filled with the target initial data, if the initial data missing on the previous target date is filled with the target initial data on the date before the previous target date, the initial data missing on the current target date may be filled with the target initial data on the date after the current target date. Namely, the missing initial filling data under each target date is filled in a bidirectional mode, so that the situation that the missing initial data under the initial date does not have the initial data of the adjacent date to be filled is avoided, and the authenticity of the filled initial data is improved.
For example, if the initial data No. 1 is used to fill the missing initial data No. 2, when there is a missing initial data No. 5 and neither the initial data No. 4 nor the initial data No. 6 is missing, for this case, the terminal device may fill the missing initial data No. 5 with the initial data No. 6 this time, and so on.
S205, if there is no initial data whose date is adjacent to the target date in the plurality of initial data corresponding to the second evaluation index, filling missing data in the plurality of initial data corresponding to the second evaluation index with the average value of all the initial data corresponding to the second evaluation index.
In the application, as can be seen from the explanation of S204, after the processing of S203-S204, there may be missing initial data, and based on this, the terminal device may use the average value of all the initial data corresponding to the second evaluation index to fill the missing initial data at the target date.
For example, if the number 3 is the target date, and there is initial data missing between the numbers 2 and 4 adjacent to the target date, the average value may be calculated and filled in for the initial data missing from the number 3.
It should be particularly noted that, when filling the number 3 missing initial data, the mean value of all the initial data corresponding to the second evaluation index generated by the object to be evaluated may be adopted; or filling the average value of all data corresponding to the second evaluation index in other target objects (target objects consistent with the preset evaluation type of the object to be evaluated). Specifically, the average value of all data corresponding to all second evaluation indexes generated by other target objects on the target date (No. 3) may be used for filling, and this is not limited. Thereby, the authenticity of filling in missing initial data can be further improved.
The target object and the object to be evaluated belong to the same field. For example, if the object to be evaluated is a company in the financial field, the target object may be a company in the financial field, and the target object is not limited to this.
And S206, determining the initial data corresponding to all the filled evaluation indexes as first target data.
In the present embodiment, after the initial data of each evaluation index pair is filled, all the initial data should be determined as the first target data. That is, the first target data includes initial data corresponding to each evaluation index that is originally not missing, and also includes initial data corresponding to each filled evaluation index.
In other embodiments, the terminal device may use a mode or a median of all non-missing initial data corresponding to each evaluation index as the initial data corresponding to the missing evaluation index, which is not limited in this application.
It should be noted that, after the first target data corresponding to each evaluation index is obtained, the first target data may be unreasonably filled, or the data generated by the object to be evaluated itself may be unreasonable. For this situation, the terminal device needs to check the first target data to further delete the abnormal first target data. Specifically, the details are as follows:
for a plurality of first target data corresponding to any evaluation index, the terminal device may perform normalization processing on each first target data respectively to obtain a plurality of normalized data; calculating the difference between each normalized data and the first median in the plurality of normalized data; determining a product value of the second median of the plurality of difference values and a preset value; generating a target interval according to the first median and the product value; and determining the standardized data belonging to the target interval as final first target data.
In application, the normalizing the first target data is to normalize the first target data corresponding to a plurality of different evaluation indexes, so that the dimensions of the first target data corresponding to the different evaluation indexes are the same.
The normalization method of the first target data corresponding to any evaluation index may be:
wherein, X k Indicating the kth first target data in all the first target data corresponding to the evaluation index; x' k Normalized data representing the kth first target data; σ may represent a standard deviation in all the first target data corresponding to the evaluation index; μ may represent a mean value among all the first target data corresponding to the evaluation index.
In an application, the first median is a median of the plurality of normalized data. The second median is a median of differences between the plurality of first normalized data and the first median, respectively. The preset value may be set according to actual conditions, and in this embodiment, the preset value is usually a constant value, and is generally equal to 1.483.
When the target interval is generated, the terminal device may perform addition and subtraction processing on the first median and the product value respectively to obtain two values, and the two values are used as the maximum value and the minimum value of the target interval. Then, the normalized data belonging to the target interval are all determined as normal first target data. And normalized data that does not belong to the target interval may be considered abnormal data.
It should be noted that, when the target interval is generated, the first median of the normalized data and the second median of the plurality of difference values are generated, so that the manner of detecting the abnormal data by the terminal device may be more reasonable. If the average value of the first target data is used for detection, the average value is easily affected by a maximum value or a minimum value in the first target data, so that the accuracy of detecting the abnormal first target data by the terminal device is low.
In application, the first score change value is used for describing the score change of the object to be evaluated in the target historical time period. Specifically, the terminal device may input the first target data into a preset scoring model for processing to obtain a first scoring change value; the preset scoring model is as follows:
wherein j represents the jth object to be evaluated; r is j Representing a first grading change value of a jth object to be evaluated; n represents the number of evaluation indexes corresponding to the object to be evaluated; x ji Representing first target data corresponding to the ith evaluation index of the jth object to be evaluated; f. of i Indicating the index weight corresponding to the ith evaluation index; mu.s j And representing the preset scoring residual error of the jth object to be evaluated.
In application, when only one object to be evaluated is present, j is 1. It can be understood that the terminal device may process a plurality of objects to be evaluated at the same time, and then generate a first score change value for each object to be evaluated.
In application, the evaluation index is different according to the type of the object to be evaluated and the preset evaluation type. For example, an evaluation index for which the object to be evaluated is a company should be different from an evaluation index employed when the object to be evaluated is a person. And when the object to be evaluated is a company, the preset evaluation type is an evaluation index of the operation risk, and the preset evaluation type is different from the evaluation index of the environment, which is not limited herein.
Specifically, the evaluation index can be generally roughly classified into three categories, namely, a national index, an industrial index and a style index. Different combinations of evaluation indexes can be selected for different preset evaluation types and objects to be evaluated. That is, the selection of the evaluation index needs to be matched with a system of a preset evaluation type.
For example, when the object to be evaluated is a company, the style index may include basic company evaluation indexes such as total assets, total liabilities, non-liquidity liabilities, total income, net profit, and the like. For the national indexes, the national indexes can comprise evaluation indexes such as annual deposit interest rate, decade national debt interest rate and the like. For the industry index, the index can be divided according to the main business of the company. Typically, for a listed company, industry metrics may be determined based on more authoritative case-level industry standards.
As can be understood from the preset scoring model, if the target historical time period is divided into a plurality of preset sampling periods, the terminal device acquires the first target data corresponding to each evaluation index in each preset sampling period. For example, when the preset sampling period is every day, the terminal device may acquire first target data corresponding to each evaluation index every day. Based on this, when calculation is performed according to the preset scoring model, the terminal device may obtain first scoring change values corresponding to the object to be evaluated in each preset sampling period.
In application, the preset scoring model can be developed by arbitrage pricing theory and capital asset pricing model. Wherein, different evaluation indexes represent the interpretation variables of the object to be evaluated when the object develops under different risk types. And the preset grading model quantitatively describes the load between each evaluation index of the object to be evaluated and a single first target datum and the linear relation between the evaluation indexes and the single first target datum on the index weight.
It should be noted that, for the same index, different experts often give different weights, so that the final obtained evaluation is often not accurate enough. Therefore, in order to obtain index weights of different suitable evaluation indexes, the terminal device may train the index weight corresponding to each evaluation index in the preset scoring model, as shown in fig. S301-S305, so as to obtain the index weight. The details are as follows:
s301, aiming at any one target object, obtaining a target score change value of the target object and second target data corresponding to each evaluation index of the target object.
In application, the target score variation value may be determined in advance by a third party company according to the second target data of the target object. The second target data may also be obtained by acquiring the target object in the manner of S201-S205, which will not be described in detail.
S302, inputting second target data corresponding to each evaluation index of the target object into the initial model for processing to obtain a prediction score change value corresponding to each evaluation index of the target object.
In application, the initial model is a model to be trained, and a formula corresponding to the initial model is similar to a formula corresponding to a preset scoring model, except that the index weight corresponding to each evaluation index in the initial model is not trained yet. Therefore, the predicted score change value may be obtained by calculating the second target data corresponding to each evaluation index through the initial model.
And S303, inputting all the prediction score change values and the target score change values of the target object into a preset loss function, and calculating the loss value of the target object. The default loss function is:
J(f)＝(r-r') T W(r-r')；
wherein J (f) represents a loss value; f represents the index weight corresponding to each evaluation index; r represents a target score change value; r' represents an array consisting of all the predicted score change values of the target object; w represents an array consisting of a plurality of preset regression weights for the target object.
In application, the initial model needs to perform loss value calculation based on the predicted score change value and the actual target score change value to obtain the loss value in the training process. And then, iterating the index weight of the initial model in the current iteration process according to the loss value.
The formula for calculating the loss value is specifically the preset loss function. It should be noted that, in an iterative process, if there are multiple target objects, the loss value of multiple target objects will be generated in the iterative process. However, in the iterative process, the iteration should be performed according to the minimum value among the loss values of all the target objects.
Specifically, in the iterative process, in step S304, if the minimum value of the loss values of all the target objects is greater than or equal to the preset loss threshold, the index weight of the initial model is iterated according to the minimum value until the minimum value is less than the preset loss threshold.
S305, if the minimum value of the loss values of all the target objects is smaller than a preset loss threshold value, generating a scoring model according to the index weight in the current iteration process.
The preset loss threshold may be set by a third-party company according to actual conditions, which is not limited herein. In each iteration process, the index weight in the initial model in the current iteration process should be updated according to the minimum value of the loss values of all the target objects in the current iteration process.
In application, in each iteration process, the formula for updating the index weight may be:
f＝(X T WX) -1 X T Wr；
wherein X represents a set of second target data corresponding to each evaluation index of the target object; w represents an array consisting of a plurality of preset regression weights for the target object; representing a transposed matrix; r represents a target score change value of the target object.
In application, when the updated index weight is obtained, the updated weight is f in the preset scoring model, and is used for calculating the first scoring change value.
S102, acquiring target events of all target objects in a target historical time period, and determining a second grading change value of the object to be evaluated according to the target events of all the target objects; the preset evaluation type of the target object is the same as that of the object to be evaluated.
In application, the target object may be an object used in training the preset scoring model, or may be another target object, and the preset evaluation type of the target object is the same as the preset evaluation type of the object to be evaluated. Further, in this embodiment, the target object may be: the object is the same as the field of the object to be evaluated and comprises data of a preset evaluation type. For example, when the object to be evaluated is a company in the financial field and the preset evaluation type is an operational risk, the target object may also be a company in the financial field and the data is data of the target object in terms of an empirical risk.
In application, the target event is an event that occurs in a target historical time period of a target object, and the occurrence of the event may affect an object to be evaluated. Typically, the impact may not be fully manifested by the first target data. Therefore, when evaluating the object to be evaluated, the target event also needs to be considered to influence the object to be evaluated. Namely, a second score change value of the object to be evaluated is generated based on the target event.
Specifically, the terminal device may determine the second score change value through S401-S404 as shown in fig. 4, which is detailed as follows:
s401, aiming at any one target object, acquiring a first actual score of the target object in each preset sampling period in a target historical time period.
In application, the preset sampling period may be set by a third company in advance, and is not limited thereto. Illustratively, when the target historical time period is in units of days, each preset sampling period is a period corresponding to each date. And the first actual score is the actual score of the target object in the preset sampling time period. The first actual score can also be set by the third-party company according to the actual condition of the target object in the preset sampling period.
It should be noted that the first actual score is generally the score of the target object after the target event occurs. The reason why the second score change value can be determined according to the target event is that: even if any target event does not occur, the score change value of the object to be evaluated in the current sampling period fluctuates. At this time, the score change value may be understood as a "normal score change value" or an "expected score change value". And when the target event occurs, the first actual score of the object to be evaluated in the current sampling period comprises two parts, wherein one part is a 'normal score change value', and the other part is the score change brought by the target event. Wherein, the other part of the score changes brought by the target events can be understood as: "abnormal score change value". Based on this, the first actual score may be considered as the sum of the "normal score change value" and the "abnormal score change value" of the target object at the time of transmitting the target event.
S402, obtaining second actual scores of the object to be evaluated in each preset sampling period in the target historical time period.
In application, the second actual score may be considered as an actual score when the object to be evaluated itself has some events in the target historical time period, or is influenced by the second target event of the target object. I.e., the second actual score is the sum of the "normal score change value" and the "abnormal score change value".
S403, aiming at any one preset sampling period, determining the simulation score of the object to be evaluated in the preset sampling period according to the mean value of the first actual scores of all the target objects in the preset sampling period.
In application, the analog score may represent a "normal score change value" of the object to be evaluated in the current sampling period. The field of the object to be detected and the field of the target object are the same as the preset evaluation type. Therefore, the score change of the object to be evaluated in the preset sampling period can be considered to have a certain relation with the score change of other target objects in the same preset sampling period.
Specifically, the terminal device may calculate the simulation score according to the following calculation formula:
E(R) it ＝α i +β i R mt ；
wherein, E (R) it Representing the simulation score of the ith object to be evaluated in the tth preset sampling period; alpha (alpha) ("alpha") i 、β i Respectively representing the scoring relation coefficients corresponding to the ith object to be evaluated; r mt And the average score change value represents the second actual score of each target object in the t preset sampling period.
Wherein, the score relation coefficient may be: the terminal device performs linear regression calculation on the second actual score of the object to be evaluated and the first actual score of the other object in advance through other historical time periods when the object to be evaluated does not have an event with an influence or is not influenced by the event of the other object. That is, linear regression is performed according to the plurality of second actual scores and the first actual scores to obtain the expression.
It should be noted that, when the object to be evaluated has no event having an influence or is not influenced by an event occurring in another target object, the first actual score should only be included as the "normal score variation value". Therefore, the score relation coefficient calculated based on the first actual score and the second actual score in the other historical time periods can be used for representing the linear relation between the object to be detected and the first actual score of the other target object when the event with influence does not occur or is not influenced by the event occurring in the other target object.
Based on this, after the score relation coefficient is obtained, the terminal device may combine all the first actual scores of all the preset sampling periods in the target historical time period to perform processing, and simulate the simulation score of the object to be detected in each preset sampling period in the target historical time period. At this point, the simulation score should be expressed as: and when the object to be detected is not influenced by the occurrence of the event with the influence or the target event of other target objects, the normal score change value of each preset sampling period is obtained.
S404, determining a second score change value according to the second actual score and the simulation score of the object to be evaluated in all the preset sampling periods.
In the application, according to the above explanation of the second actual score in S402 and the explanation of the analog score corresponding to S403, the terminal device may use the difference between the first actual score and the analog score as the second score change value when the object to be detected has an influence in the preset sampling period or is influenced by the target event.
However, it should be noted that when the second score change value is obtained, it is further determined whether the influence of the target event to the object to be detected reaches a preset requirement according to the second score change value, so as to further determine whether the object to be detected needs to be evaluated based on the second score change value.
Specifically, the terminal device may calculate an average score change value of the object to be evaluated in the target historical time period according to the second actual scores and the simulated scores of the object to be evaluated in all the preset sampling time periods. And then, determining whether the target event belongs to a preset influence event of the object to be evaluated according to the average score change value.
The calculation of the average score change value of the object to be evaluated in the target historical time period may be regarded as that a difference value between the second actual score and the simulation score in each preset sampling time period is used as the second score change value in the preset sampling time period. And then, calculating the mean value of the second score change values of all the objects to be evaluated to obtain the average score change value of the objects to be evaluated in the target historical time period.
Specifically, the calculation formula may be:
wherein CAR represents the mean score change value; n represents the number of differences; y represents the number of preset sampling periods; AR xy Representing the x-th difference at the y-th preset sampling period.
It can be understood that if the average score change value is lower than the preset change value, it may be determined that the target event does not belong to the preset influence event of the object to be evaluated. For example, if the average score variation value is 0, it may be considered that the simulated score of the object to be evaluated is consistent with the second actual score. At this time, it may be considered that the target event does not affect the object to be detected. If the average score change value is greater than or equal to the preset change value, the target event can be determined to belong to the preset influence event of the object to be evaluated. I.e. the target event will have an impact on the object to be detected.
In an embodiment, the preset variation value may be set according to an actual situation, and is not limited thereto. The preset influence event is an event having a significant influence on the development of the event to be detected, and the influence degree exceeds the influence degree corresponding to the preset change value.
It should be noted that, when it is determined that the target event does not belong to the preset influence event of the object to be evaluated, the terminal device may determine the second score change value according to preset values of all preset sampling periods. The preset value may be 0. That is, when it is determined that the target event does not affect the data to be evaluated, it may be considered that the influence of the second score change value on the object to be evaluated is not required to be considered. Therefore, at the time of calculation, the subsequent processing may be performed with the second score change value as 0.
However, when the target event belongs to a preset influence event of the object to be evaluated, the terminal device may determine a difference value between the second actual score and the analog score of each preset sampling period as a candidate score change value of each preset sampling period, and determine a set of candidate score change values of all preset sampling periods as the second score change value.
Wherein each candidate variation value is: and under each preset sampling period, the change degree of the object to be detected when the object to be detected is influenced by the occurrence of the target event. According to the explanation of the second actual score in S402 and the explanation of the simulated score corresponding to S403, the candidate score variation value obtained by subtracting the second actual score from the simulated score is the "abnormal score variation value".
Based on this, in this embodiment, after the second score change value is obtained, the terminal device may further determine whether the object to be evaluated needs to be evaluated in combination with the second score change value, instead of directly evaluating in combination, according to the average score change value of the object to be evaluated in the target historical time period, so as to further improve the accuracy of evaluating the object to be evaluated.
S103, determining a target score value of the object to be evaluated in the target historical time period according to the first score change value and the second score change value.
In an application, since the target historical time period includes a plurality of preset sampling periods, the target score value also typically includes a sampling score value corresponding to each preset sampling period.
Specifically, when calculating the sampling score value corresponding to each preset sampling period, the terminal device may first obtain an initial score value of the object to be evaluated; determining a total score change value of an object to be evaluated in each preset sampling period according to the first score change value and the second score change value of each preset sampling period in the target historical time period; inputting the total score change value of the object to be evaluated in each preset sampling period into a preset recurrence formula for processing to obtain the sampling score value of the object to be evaluated in each preset sampling period; the preset recurrence formula is as follows:
P t ＝P 0 ×(1+r 1 )×(1+r 2 )×...×(1+r t )；
wherein, P t Representing the sampling score value of the object to be evaluated in the t-th preset sampling period; p 0 Representing an initial value of credit; r is t Representing the total score change value of the object to be evaluated in the t-th preset sampling period; and determining a target score value according to the sampling score values of the object to be evaluated in all the preset sampling time periods.
In application, the initial score value is generally the score of the object to be evaluated in the first preset sampling period in the target history period. The initial value of credit may be set by a third party company, but is not limited thereto.
In an application, for the total score change value of each preset sampling period, it may be a sum of the first score change value and the second score change value of each preset sampling period. It should be noted that, as can be seen from the above explanation of S404, when it is determined that the target event does not belong to the preset influence event of the object to be evaluated, or when the object to be evaluated has no influence event, it may be considered that the terminal device does not need to consider the influence of the second score change value on the object to be evaluated. At this time, the total score variation value is the first score variation value. If the target event is determined to belong to a preset influence event of the object to be evaluated, or the object to be evaluated has an influence event, the total score change value can be considered as the sum of the first score change value and the second score change value.
When a target event occurs, the influence of the event to be evaluated should be continuous, and therefore, the above recursive formula needs to be adopted to obtain the sampling score value of the object to be evaluated in each preset sampling period. Then, the terminal device may determine a final target score value according to all the sampling score values. For example, the sum of all the sample score values is taken as the final target score value. And then, the terminal equipment can output the target evaluation of the object to be evaluated according to the target evaluation value. For example, the terminal device may store a score interval corresponding to each target evaluation in advance. And then, outputting corresponding target evaluation according to the grading interval to which the target grading value belongs.
In the embodiment, first target data of the object to be evaluated in the target history time period and target events of other target objects belonging to the same evaluation type in the target history time period are respectively acquired according to a preset evaluation type of the object to be evaluated. Then, the terminal device may determine a first score change value of the object to be evaluated according to the first target data. And meanwhile, determining a second grading change value of the object to be evaluated in the target historical time period according to the target event. Finally, the terminal device may generate a target score value of the object to be evaluated in the target window period based on the first score change value and the second score change value at the same time. Therefore, when the target scoring value is generated, not only the first target data of the object to be evaluated is only based, but also the influence possibly generated by each target event of the target objects with the same evaluation type in the target historical time period is considered. Furthermore, the target scoring value of the company can be obtained from multiple aspects, so that the finally obtained target scoring value is more objective and accurate
It should be added that, in S102, when the terminal device acquires target events that occur in the target history time period of each target object, the terminal device typically acquires all events that occur in the target history time period of each target object. However, all events are not related to a preset evaluation type and are not events that can affect the object to be evaluated. Therefore, the terminal device further needs to filter all the acquired events through S501-S508 shown in fig. 5. The details are as follows:
s501, acquiring a plurality of first events which occur in the target historical time period and are consistent with the evaluation types of the target objects.
And S502, respectively determining the event type of each first event.
In application, the first event is an event which occurs in a target historical time period and is consistent with a preset evaluation type of the target object. For example, if the preset evaluation type is operation risk, the first event may be an operation-related event. For example, XX company has a product problem, resulting in an event that the business situation is problematic.
In application, the event types are generally classified into numerical events belonging to a numerical type, and text events belonging to a text type. In the case of a numerical event, it may be considered that the main information included in the event is embodied in the form of a numerical value. For text-type events, it can be considered that the contained main information is embodied in the form of words.
Generally, the event type of each first event may be determined by an event classification model preset in the terminal device, or may be set by a third-party company, and further, the terminal device may directly determine the event type of each first event, which is not limited herein.
S503, determining whether all numerical events belong to preset influence events of the object to be evaluated according to numerical values contained in all numerical events aiming at all numerical events of which the event types belong to numerical types.
In application, when determining whether all numerical events belong to the preset influence events of the object to be evaluated according to the numerical values included in all the numerical events, the determination may be performed according to the mean value of all the numerical values, or may be performed according to the standard deviation of all the numerical values. For example, when the mean value or the standard deviation is smaller than the preset influence value, it may be determined that all numerical events do not belong to the preset influence event of the object to be evaluated; otherwise, determining that all numerical events belong to the preset influence events of the object to be evaluated. In this embodiment, the manner of determining whether all the numerical events belong to the preset influence events of the object to be evaluated is not limited.
And S504, if all the numerical events are determined not to belong to the preset influence events of the object to be evaluated, sequentially and alternately deleting the numerical events corresponding to the maximum value and the minimum value in all the numerical values until all the numerical events are determined to belong to the preset influence events of the object to be evaluated.
And S505, if all the numerical events are determined to belong to the preset influence events of the object to be evaluated, determining all the numerical events which are not deleted as target events.
In application, the above alternate deletion may be understood as that, when the last deleted numerical event is the numerical event corresponding to the maximum value, the numerical event corresponding to the minimum value may be deleted at present. When an event is deleted once, whether all the numerical events which are not deleted belong to the preset influence events of the object to be evaluated or not should be determined again. Therefore, the step S504 or S505 is executed again until all the numerical events are determined to belong to the preset influence events of the object to be evaluated, and all the numerical events which are not deleted are determined as the target events.
S506, aiming at all text events of which all event types belong to character types, acquiring second events which occur in a target historical time period and are consistent with the evaluation types of the objects to be evaluated; the event type of the second event is a character type.
In application, the second event is an event which occurs in the target historical time period of the object to be evaluated and is consistent with the evaluation type. That is, the second event may be considered as an event having an influence on the object to be evaluated, and the main information included in the second event is also embodied in the form of characters.
And S507, respectively calculating the similarity between each text event and the second event.
And S508, determining the text event with the similarity larger than the preset similarity as the target event.
In application, the similarity can be determined by a similarity model. Specifically, the similarity model may represent each text event and the second event as a numerical vector that may be calculated by the terminal device, that is, in the form of a feature. Then, the similarity between the two is calculated according to the numerical value vector. For example, the euclidean distance or the cosine distance of both are calculated from the numerical vectors to obtain the similarity. In this embodiment, the calculation method of the similarity is not limited.
In application, the preset similarity may be set by a third-party company according to actual conditions, which is not limited.
It should be noted that, for the target events obtained in S505 and S508, if the total number of all the target events is lower than the preset number, for example, lower than 10, it may be considered that the obtained target events are too few to better reflect the influence of the target events on the object to be evaluated. Based on this, when the number of target events is lower than the preset number, each of the plurality of first events acquired in S501 may be directly determined as a target event.
However, it should be particularly noted that, when obtaining the target event, the influence degree of the target event on the object to be evaluated may be small due to the large scale of the number of the target events, but when the terminal device determines whether the target event has influence according to the average score change value of all the second score change values, the terminal device may also consider that there is a certain influence.
Based on this, in order to accurately determine whether all target events belong to the preset influence events of the object to be evaluated, the following steps are described in the above S404: after the average score change value is lower than the preset change value and the target event is determined not to belong to the preset influence event of the object to be evaluated, the terminal device can introduce the concept of the preset check value to further check the target event.
Specifically, the terminal device may calculate a standard deviation of the object to be evaluated in the target historical time period according to the second actual score and the simulation score in each preset sampling time period. Then, distribution values of the mean score change value and the standard deviation when the target distribution is obeyed are determined, and a ratio of the mean score change value and the standard deviation is calculated. Then, determining a check value according to the difference between the distribution value and the ratio; the check value is used for checking whether the target event belongs to a preset influence event of the object to be evaluated. If the check value is larger than the preset check value, determining that the target event belongs to a preset influence event of the object to be evaluated; if the check value is less than or equal to the check value, determining that the target event does not belong to the preset influence event of the object to be evaluated
The formula for calculating the standard deviation is the prior art, and will not be described in detail. It should be noted that, after the standard deviation is obtained, the terminal device may determine, according to the average score change value and the standard deviation, the target distribution obeyed by the average score change value and the standard deviation, and further obtain a corresponding distribution value.
The target distribution includes, but is not limited to, various types such as normal distribution, binomial distribution, bernoulli distribution, t-distribution, and the like. In this embodiment, the specific type of the target distribution is not limited. For example, in the present embodiment, the target distribution may be specifically a t-distribution, which may be used to estimate a mean of a population that is normally distributed and whose variance is unknown from a small sample. Wherein, the distribution value is the t value of the average score change value and the standard deviation when obeying the t distribution.
And the check value is a specific numerical value correspondingly determined according to the ratio. For example, the terminal device may be preset with a plurality of difference intervals, and each difference interval corresponds to one probability value. And then, determining a final check value according to the difference interval where the difference is located and the probability value corresponding to the difference interval.
Illustratively, the check value is determined as follows:
power＝P(Z>t-d)+1-P(Z>-t-d)；
wherein power is a check value; t is the above distribution value; d is the above ratio; z > t-d represents a difference interval with a value larger than t-d in a plurality of preset difference intervals; p (Z > t-d) represents the sum of all probability values corresponding to the difference interval with the numerical value larger than t-d.
Finally, after the check value is obtained, the check value can be compared with a preset check value, and then the target event is determined to belong to the preset influence event of the object to be evaluated. The preset check value can be set by a third-party company according to actual conditions. For example, the preset check value may be 0.8.
Referring to fig. 6, fig. 6 is a block diagram of an object evaluation apparatus according to an embodiment of the present disclosure. The object evaluation apparatus in this embodiment includes modules for executing the steps in the embodiments corresponding to fig. 1 to 5. Please specifically refer to fig. 1 to 5 and related descriptions of embodiments corresponding to fig. 1 to 5. For convenience of explanation, only the portions related to the present embodiment are shown. Referring to fig. 6, the object evaluation apparatus 600 may include: a first score change value determination module 610, a second score change value determination module 620, and a target score value determination module 630, wherein:
the first score change value determining module 610 is configured to obtain first target data of the object to be evaluated in a target historical time period according to a preset evaluation type of the object to be evaluated, and determine a first score change value of the object to be evaluated according to the first target data.
The second score change value determining module 620 is configured to obtain target events of each target object occurring within a target historical time period, and determine a second score change value of the object to be evaluated according to the target events of all the target objects; the preset evaluation type of the target object is the same as that of the object to be evaluated.
And a target score value determining module 630, configured to determine a target score value of the object to be evaluated in the target historical time period according to the first score change value and the second score change value.
In one embodiment, the first score change value determination module 610 is further configured to:
acquiring a plurality of initial data corresponding to a plurality of evaluation indexes of an object to be evaluated in a target historical time period; if a plurality of initial data corresponding to a first evaluation index in the plurality of evaluation indexes are missing, filling missing data in the plurality of initial data corresponding to the first evaluation index with a target value; if a plurality of initial data corresponding to a second evaluation index in the plurality of evaluation indexes are missing, determining a target date of the missing data in the plurality of initial data corresponding to the second evaluation index; if target initial data with a date adjacent to the target date exists in the plurality of initial data corresponding to the second evaluation index, filling missing data in the plurality of initial data corresponding to the second evaluation index by adopting the target initial data; if the initial data with the date adjacent to the target date does not exist in the plurality of initial data corresponding to the second evaluation index, filling missing data in the plurality of initial data corresponding to the second evaluation index by adopting the average value of all the initial data corresponding to the second evaluation index; and determining initial data corresponding to all the filled evaluation indexes as first target data.
In one embodiment, the first score change value determination module 610 is further configured to:
inputting the first target data into a preset grading model for processing to obtain a first grading change value; the preset scoring model is as follows:
wherein j represents the j th object to be evaluated; r is j Representing a first grading change value of a jth object to be evaluated; n represents the number of evaluation indexes corresponding to the object to be evaluated; x ji Representing first target data corresponding to the ith evaluation index of the jth object to be evaluated; f. of i Indicating index weight corresponding to the ith evaluation index; mu.s j Representing the jth object to be evaluatedAnd presetting scoring residual errors.
In one embodiment, the object evaluation device 600 further comprises:
and the second target data acquisition module is used for acquiring a target score change value of the target object and second target data corresponding to each evaluation index of the target object aiming at any one target object.
And the prediction score change value generation module is used for inputting the second target data corresponding to each evaluation index of the target object into the initial model for processing to obtain the prediction score change value corresponding to each evaluation index of the target object.
The loss value generation module is used for inputting all the predicted score change values and the target score change values of the target object into a preset loss function and calculating the loss value of the target object; the default loss function is:
J(f)＝(r-r') T W(r-r')；
wherein J (f) represents a loss value; f represents the index weight corresponding to each evaluation index; r represents a target score change value; r' represents an array consisting of all the predicted score change values of the target object; w represents an array consisting of a plurality of preset regression weights for the target object.
And the iteration module is used for iterating the index weight of the initial model according to the minimum value if the minimum value of the loss values of all the target objects is greater than or equal to a preset loss threshold value until the minimum value is less than the preset loss threshold value.
And the scoring model generating module is used for generating a scoring model according to the index weight in the current iteration process if the minimum value of the loss values of all the target objects is smaller than a preset loss threshold value.
In an embodiment, the second score change value determination module 620 is further configured to:
aiming at any one target object, acquiring a first actual score of the target object in each preset sampling period in a target historical time period; acquiring second actual scores of the object to be evaluated in each preset sampling period in the target historical time period; aiming at any one preset sampling period, determining the simulation score of the object to be evaluated in the preset sampling period according to the mean value of the first actual scores of all the target objects in the preset sampling period; and determining a second score change value according to the second actual score and the simulation score of the object to be evaluated in all the preset sampling periods.
In an embodiment, the second score change value determining module 620 is further configured to:
calculating an average score change value of the object to be evaluated in a target historical time period according to the second actual scores and the simulated scores of the object to be evaluated in all preset sampling time periods; determining whether the target event belongs to a preset influence event of the object to be evaluated or not according to the average grading change value; if the target event does not belong to the preset influence event of the object to be evaluated, determining a second grading change value according to preset values of all preset sampling time periods; and if the target event belongs to the preset influence events of the object to be evaluated, determining the difference value between the second actual score and the simulation score of each preset sampling time period as the candidate score change value of each preset sampling time period, and determining the set of the candidate score change values of all the preset sampling time periods as the second score change value.
In an embodiment, the target score value determination module 630 is further configured to:
acquiring an initial score value of an object to be evaluated; determining a total score change value of an object to be evaluated in each preset sampling period according to the first score change value and the second score change value of each preset sampling period in the target historical time period; inputting the total grade change value of the object to be evaluated in each preset sampling period into a preset recurrence formula for processing to obtain the sampling grade value of the object to be evaluated in each preset sampling period; the preset recurrence formula is as follows:
P t ＝P 0 ×(1+r 1 )×(1+r 2 )×...×(1+r t )；
wherein, P t Representing the sampling score value of the object to be evaluated in the t-th preset sampling period; p 0 Representing an initial value of credit; r is t Representing the total score change value of the object to be evaluated in the t-th preset sampling period; according to the object to be evaluated in all the testsAnd setting the sampling score value of the sampling time period, and determining a target score value.
It should be understood that, in the structural block diagram of the object evaluation apparatus shown in fig. 6, each module is used to execute each step in the embodiment corresponding to fig. 1 to fig. 5, and each step in the embodiment corresponding to fig. 1 to fig. 5 has been explained in detail in the above embodiment, and specific reference is made to the relevant description in the embodiments corresponding to fig. 1 to fig. 5 and fig. 1 to fig. 5, which is not repeated herein.
Fig. 7 is a block diagram of a terminal device according to an embodiment of the present application. As shown in fig. 7, the terminal device 700 of this embodiment includes: a processor 710, a memory 720, and a computer program 730, such as a program for an object evaluation method, stored in the memory 720 and executable on the processor 710. The processor 710, when executing the computer program 730, implements the steps in the embodiments of the object evaluation methods described above, such as S101 to S103 shown in fig. 1. Alternatively, the processor 710, when executing the computer program 730, implements the functions of the modules in the embodiment corresponding to fig. 6, for example, the functions of the modules 610 to 630 shown in fig. 6, and refer to the related description in the embodiment corresponding to fig. 6 specifically.
Illustratively, the computer program 730 may be divided into one or more modules, and the one or more modules are stored in the memory 720 and executed by the processor 710 to implement the object evaluation method provided by the embodiments of the present application. One or more of the modules may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution of the computer program 730 in the terminal device 700. For example, the computer program 730 may implement the object evaluation method provided by the embodiment of the present application.
The processor 710 may be a central processing unit, but may also be other general purpose processors, digital signal processors, application specific integrated circuits, off-the-shelf programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The storage 720 may be an internal storage unit of the terminal device 700, such as a hard disk or a memory of the terminal device 700. The memory 720 may also be an external storage device of the terminal device 700, such as a plug-in hard disk, a smart card, a flash memory card, etc. provided on the terminal device 700. Further, the memory 720 may also include both internal and external memory units of the terminal device 700.
The embodiment of the present application provides a computer-readable storage medium, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the object evaluation method in the above embodiments.
The embodiment of the present application provides a computer program product, which, when running on a terminal device, enables the terminal device to execute the object evaluation method in the foregoing embodiments.
The above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present application and are intended to be included within the scope of the present application.
Claims (10)
1. A method for evaluating an object, the method comprising:
according to the preset evaluation type of the object to be evaluated, first target data of the object to be evaluated in a target historical time period are obtained, and a first score change value of the object to be evaluated is determined according to the first target data;
acquiring target events of all target objects in a target historical time period, and determining a second grading change value of the object to be evaluated according to the target events of all the target objects; the preset evaluation type of the target object is the same as that of the object to be evaluated;
and determining the target score value of the object to be evaluated in the target historical time period according to the first score change value and the second score change value.
2. The method according to claim 1, wherein obtaining first target data of the object to be evaluated within a target history time period comprises:
acquiring a plurality of initial data corresponding to a plurality of evaluation indexes of an object to be evaluated in a target historical time period;
if a plurality of initial data corresponding to a first evaluation index in the plurality of evaluation indexes are missing, filling missing data in the plurality of initial data corresponding to the first evaluation index with a target value;
if a plurality of initial data corresponding to a second evaluation index in the plurality of evaluation indexes are missing, determining a target date of the missing data in the plurality of initial data corresponding to the second evaluation index;
if target initial data with a date adjacent to the target date exists in the plurality of initial data corresponding to the second evaluation index, filling missing data in the plurality of initial data corresponding to the second evaluation index by adopting the target initial data;
if the initial data with the date adjacent to the target date does not exist in the plurality of initial data corresponding to the second evaluation index, filling missing data in the plurality of initial data corresponding to the second evaluation index by adopting the average value of all the initial data corresponding to the second evaluation index;
and determining initial data corresponding to all the filled evaluation indexes as first target data.
3. The method according to claim 1, wherein determining a first score change value of an object to be evaluated based on the first objective data comprises:
inputting the first target data into a preset grading model for processing to obtain a first grading change value; the preset scoring model is as follows:
wherein j represents the j th object to be evaluated; r is j Representing a first grading change value of a jth object to be evaluated; n represents the number of evaluation indexes corresponding to the object to be evaluated; x ji Representing first target data corresponding to the ith evaluation index of the jth object to be evaluated; f. of i Indicating the index weight corresponding to the ith evaluation index; mu.s j And representing the preset scoring residual error of the jth object to be evaluated.
4. The method of claim 3, further comprising, prior to inputting the first objective data into the predetermined scoring model for processing:
aiming at any one target object, acquiring a target score change value of the target object and second target data corresponding to each evaluation index of the target object;
inputting second target data corresponding to each evaluation index of the target object into the initial model for processing to obtain a prediction score change value corresponding to each evaluation index of the target object;
inputting all the prediction score change values and the target score change values of the target object into a preset loss function, and calculating the loss value of the target object; the default loss function is:
J(f)＝(r-r') T W(r-r')；
wherein J (f) represents a loss value; f represents index weight corresponding to each evaluation index; r represents a target score change value; r' represents an array consisting of all the predicted score change values of the target object; w represents an array consisting of a plurality of preset regression weights for the target object;
if the minimum value of the loss values of all the target objects is larger than or equal to a preset loss threshold value, iterating the index weight of the initial model according to the minimum value until the minimum value is smaller than the preset loss threshold value;
and if the minimum value of the loss values of all the target objects is smaller than a preset loss threshold value, generating a grading model according to the index weight in the current iteration process.
5. The method according to claim 1, wherein the step of obtaining target events of each target object occurring within a target historical time period and determining a second score change value of the object to be evaluated according to the target events of all the target objects comprises:
aiming at any one target object, acquiring a first actual score of the target object in each preset sampling period in a target historical time period;
acquiring second actual scores of the object to be evaluated in each preset sampling period in the target historical time period;
aiming at any one preset sampling period, determining the simulation score of the object to be evaluated in the preset sampling period according to the mean value of the first actual scores of all the target objects in the preset sampling period;
and determining a second score change value according to the second actual score and the simulation score of the object to be evaluated in all the preset sampling periods.
6. The method according to claim 5, wherein determining a second score change value according to the second actual score and the simulated score of the object to be evaluated in all the preset sampling periods comprises:
calculating an average score change value of the object to be evaluated in a target historical time period according to the second actual scores and the simulated scores of the object to be evaluated in all the preset sampling time periods;
determining whether the target event belongs to a preset influence event of the object to be evaluated or not according to the average score change value;
if the target event does not belong to the preset influence event of the object to be evaluated, determining a second grading change value according to preset values of all preset sampling time periods;
and if the target event belongs to the preset influence events of the object to be evaluated, determining the difference value between the second actual score and the simulation score of each preset sampling time period as the candidate score change value of each preset sampling time period, and determining the set of the candidate score change values of all the preset sampling time periods as the second score change value.
7. The method according to any one of claims 1 to 6, wherein determining a target score value of the object to be evaluated in the target historical time period according to the first score change value and the second score change value comprises:
acquiring an initial score value of an object to be evaluated;
determining a total score change value of an object to be evaluated in each preset sampling period according to the first score change value and the second score change value of each preset sampling period in the target historical time period;
inputting the total score change value of the object to be evaluated in each preset sampling period into a preset recurrence formula for processing to obtain the sampling score value of the object to be evaluated in each preset sampling period; the preset recurrence formula is as follows:
P t ＝P 0 ×(1+r 1 )×(1+r 2 )×...×(1+r t )；
wherein, P t The method comprises the steps of representing the sampling score value of an object to be evaluated in the t-th preset sampling period; p 0 Representing an initial value of credit; r is t Representing the total score change value of the object to be evaluated in the t-th preset sampling period;
and determining a target score value according to the sampling score values of the object to be evaluated in all the preset sampling periods.
8. An object evaluation apparatus, characterized in that the apparatus comprises:
the first grading change value determining module is used for acquiring first target data of the object to be evaluated in a target historical time period according to the preset evaluation type of the object to be evaluated and determining a first grading change value of the object to be evaluated according to the first target data;
the second score value determining module is used for acquiring target events of all target objects in a target historical time period and determining second score change values of the objects to be evaluated according to the target events of all the target objects; the preset evaluation type of the target object is the same as that of the object to be evaluated;
and the target score value determining module is used for determining the target score value of the object to be evaluated in the target historical time period according to the first score change value and the second score change value.
9. A terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the method according to any of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211019789.6A CN115860505A (en) | 2022-08-24 | 2022-08-24 | Object evaluation method and device, terminal equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211019789.6A CN115860505A (en) | 2022-08-24 | 2022-08-24 | Object evaluation method and device, terminal equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115860505A true CN115860505A (en) | 2023-03-28 |
Family
ID=85660631
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211019789.6A Pending CN115860505A (en) | 2022-08-24 | 2022-08-24 | Object evaluation method and device, terminal equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115860505A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116664018A (en) * | 2023-07-28 | 2023-08-29 | 华能济南黄台发电有限公司 | Power plant equipment running state evaluation platform |
-
2022
- 2022-08-24 CN CN202211019789.6A patent/CN115860505A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116664018A (en) * | 2023-07-28 | 2023-08-29 | 华能济南黄台发电有限公司 | Power plant equipment running state evaluation platform |
| CN116664018B (en) * | 2023-07-28 | 2023-10-31 | 华能济南黄台发电有限公司 | Power plant equipment running state evaluation platform |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI788529B (en) | Credit risk prediction method and device based on LSTM model | |
| Baudry et al. | A machine learning approach for individual claims reserving in insurance | |
| CN107025596B (en) | Risk assessment method and system | |
| Chen et al. | Bayesian forecasting for financial risk management, pre and post the global financial crisis | |
| Coles et al. | An introduction to statistical modeling of extreme values | |
| CN110544155B (en) | User credit score acquisition method, acquisition device, server and storage medium | |
| Pigeon et al. | Individual loss reserving using paid–incurred data | |
| US20110047058A1 (en) | Apparatus and method for modeling loan attributes | |
| CN112734559A (en) | Enterprise credit risk evaluation method and device and electronic equipment | |
| CN112990386B (en) | User value clustering method and device, computer equipment and storage medium | |
| Baudry et al. | Non parametric individual claim reserving in insurance | |
| Qiu | Estimation of tail risk measures in finance: Approaches to extreme value mixture modeling | |
| CN115860505A (en) | Object evaluation method and device, terminal equipment and storage medium | |
| US10878329B1 (en) | Probabilistic accumulation approach to assess primary uncertainty in catastrophe models | |
| Hamidieh | Estimating the tail shape parameter from option prices | |
| CN113988431A (en) | Method, system and equipment for predicting potential broker capacity of client | |
| BULUT et al. | Financial Failure Estimation with Logistic Regression Model: A Study on Technology Sector Companies Treated in BIST | |
| CN114784795A (en) | Wind power prediction method and device, electronic equipment and storage medium | |
| Antonio et al. | Micro–level stochastic loss reserving | |
| CN114418776A (en) | Data processing method, device, terminal equipment and medium | |
| Lu | Monte Carlo analysis of methods for extracting risk‐neutral densities with affine jump diffusions | |
| McNulty | Severity Curve Fitting for Long Tailed Lines: An Application of Stochastic Processes and Bayesian Models | |
| Souissi | The implied risk neutral density dynamics: Evidence from the S&P TSX 60 index | |
| Maniati et al. | A Neural Network approach for integrating banks’ decision in shipping finance | |
| Caja et al. | Influence of economic factors on the credit rating transitions and defaults of credit insurance business |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |