CN115759283A - Model interpretation method and device, electronic equipment and storage medium - Google Patents
Model interpretation method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN115759283A CN115759283A CN202211337770.6A CN202211337770A CN115759283A CN 115759283 A CN115759283 A CN 115759283A CN 202211337770 A CN202211337770 A CN 202211337770A CN 115759283 A CN115759283 A CN 115759283A
- Authority
- CN
- China
- Prior art keywords
- model
- influence
- graph
- interpretation
- feature
- 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 invention discloses a model interpretation method, a model interpretation device, electronic equipment and a storage medium. Wherein, the method comprises the following steps: obtaining a prediction result corresponding to the prediction sample through a financial product marketing model; generating, by a machine learning model interpretation tool, an influence global graph and at least one influence trend graph matched with the financial product marketing model, and at least one feature interpretation graph and at least one feature influence graph matched with the prediction samples; and generating a model interpretation report according to the influence global graph, the influence trend graph, the characteristic interpretation graph and the characteristic influence graph so that the user can carry out financial product marketing according to the prediction result and the model interpretation report. According to the technical scheme, the machine learning model interpretation technology is utilized, the interpretation information of the model is automatically generated without depending on the model type, and the generation efficiency of the model interpretation report is improved.
Description
Technical Field
The present invention relates to the field of data processing technologies, and in particular, to a model interpretation method and apparatus, an electronic device, and a storage medium.
Background
In the process of digital transformation in the financial industry, machine learning algorithms and artificial intelligence models are more and more widely applied. In practical application of the machine learning model, the model is trained by using mass sample data, a result is calculated by the trained model according to learning experience, and finally the result output by the model is used as a decision basis. For example, the financial product marketing model may output a credit score, a marketing success probability, and the like of the user according to the input user information.
The accuracy and the automation degree of the machine learning model are higher, however, the output result of the machine learning model is lack of traceability, and the decision process is often too abstract for a user to understand. Meanwhile, the types of the machine learning models are various, for example, the machine learning models can be based on algorithms such as logistic regression, random forest, XGBoost and the like, and therefore, a model interpretation method independent of the model type is urgently needed so that a user can use the models more fully.
Disclosure of Invention
The invention provides a model interpretation method, a model interpretation device, electronic equipment and a storage medium, which are used for automatically generating interpretation information of a model independent of the model type.
In a first aspect, an embodiment of the present invention provides a model interpretation method, where the method includes:
obtaining a prediction result corresponding to the prediction sample through a financial product marketing model;
generating, by a machine learning model interpretation tool, an influence global graph and at least one influence trend graph matched with the financial product marketing model, and at least one feature interpretation graph and at least one feature influence graph matched with the prediction samples;
and generating a model interpretation report according to the influence global graph, the influence trend graph, the characteristic interpretation graph and the characteristic influence graph so as to enable the user to carry out financial product marketing according to the prediction result and the model interpretation report.
In a second aspect, an embodiment of the present invention further provides a model interpretation apparatus, where the apparatus includes:
the prediction result acquisition module is used for acquiring a prediction result corresponding to the prediction sample through a financial product marketing model;
the model interpretation module is used for generating an influence global graph and at least one influence trend graph matched with the financial product marketing model and at least one characteristic interpretation graph and at least one characteristic influence graph matched with the prediction sample through a machine learning model interpretation tool;
and the model interpretation report generation module is used for generating a model interpretation report according to the influence global graph, the influence trend graph, the characteristic interpretation graph and the characteristic influence graph so as to enable a user to carry out financial product marketing according to the prediction result and the model interpretation report.
In a third aspect, an embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the model interpretation method according to any one of the embodiments of the present invention when executing the program.
In a fourth aspect, embodiments of the present invention also provide a storage medium storing computer-executable instructions for performing the model interpretation method according to any one of the embodiments of the present invention when executed by a computer processor.
According to the technical scheme of the embodiment of the invention, the machine learning model interpretation technology is utilized, the interpretation information of the model is automatically generated without depending on the model type, and the generation efficiency of the model interpretation report is improved.
It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present invention, nor do they necessarily limit the scope of the invention. Other features of the present invention will become apparent from the following description.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of a model interpretation method according to an embodiment of the present invention;
FIG. 2 is a global graph of influence according to an embodiment of the present invention;
FIG. 3 is a characteristic explanation diagram of the management scale of characteristic monthly-daily-average assets according to an embodiment of the invention;
FIG. 4 is a characteristic impact diagram provided by an embodiment of the invention;
FIG. 5 is another feature impact diagram provided by an embodiment of the invention;
FIG. 6 is a flowchart of a model interpretation method according to a second embodiment of the present invention;
FIG. 7 is a graph illustrating the trend of influence on the scale of characteristic monthly-daily asset management according to a second embodiment of the present invention;
fig. 8 is a schematic structural diagram of a model interpretation apparatus according to a third embodiment of the present invention;
fig. 9 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention.
Detailed Description
In order to make those skilled in the art better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in other sequences than those illustrated or described herein. Moreover, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Example one
Fig. 1 is a flowchart of a model interpretation method, which is applicable to the case of generating a model interpretation report of a financial product marketing model, according to an embodiment of the present invention, and the method may be implemented by a model interpretation apparatus, which may be implemented in hardware and/or software, and may be configured in an electronic device with data processing capability.
As shown in fig. 1, the method includes:
s110, obtaining a prediction result corresponding to the prediction sample through a financial product marketing model;
the financial product marketing model can be obtained through machine learning model training of a data mining algorithm, specifically can be logistic regression, random forest, XGboost and the like, and is not limited in the application.
The prediction sample may be some asset information corresponding to the prediction object, such as personal assets and liability conditions of the prediction object, which is not limited in this application. Inputting the prediction sample into a financial product marketing model, processing the prediction sample by the model to obtain a prediction result corresponding to the prediction sample, wherein the prediction result can be marketing success probability or personal credit rating and the like, and the method is not limited in the application.
S120, generating an influence global graph and at least one influence trend graph matched with the financial product marketing model and at least one feature interpretation graph and at least one feature influence graph matched with the prediction sample through a machine learning model interpretation tool;
wherein the machine learning model interpretation tool may be a machine learning model interpretation tool using model independent methods, model independent methods being model interpretable methods that are independent of a particular model type, as opposed to model dependent methods, which may be applied to any machine learning model.
The model-independent method is adopted in the embodiment, compared with the traditional method, the universality and the operation efficiency of the interpretation tool are greatly improved, the reliability and the transparency of the model are established, the trust of the model is established, and a user is helped to make a better marketing decision.
In the embodiment, the preferred machine learning model interpretation tool is SHAP (SHAPLey Additive explantations), which is a game theory method for interpreting the output of any machine learning model.
The SHAP value may be used by the SHAP to interpret the importance of each feature.
The SHAP value discusses the problem of how to distribute the final benefits of a group of individuals who play in cooperation or opposition to each other, and the SHAP method is determined according to the contribution of each participant, namely each feature is equivalent to one participant, the coefficient (in a linear model) of the feature is equivalent to the contribution rate, and therefore the coefficient multiplied by the average value is the average contribution value. For a single instance xi, the contribution Φ ij of the jth feature to the prediction is the contribution of Xij minus the average.
The SHAP value for the feature value j may be interpreted as the contribution Φ j of the jth feature value to the prediction of this particular instance compared to the average prediction of the dataset, the SHAP value being suitable for classification and regression analysis of the dataset. The SHAP value for each feature value is a weighted sum over all possible feature value combinations, i.e., the contribution to the total prediction:
where S is a subset of the features used in the model, x is a vector of eigenvalues of the instance to be interpreted, and p is the number of features. val (S) is a predicted value for the feature value in the set S, and is a result of marginalizing features not included in the set S.
After the financial product marketing model is trained, generating an influence global graph and at least one influence trend graph matched with the financial product marketing model through a SHAP tool.
On the basis of the above embodiment, optionally, the influence global graph is used for representing each feature of the financial product marketing model, and the influence on the prediction result changes along with the change of the relative value of the feature value.
Among other features of the financial product marketing model may include: the recent expenditure amount, the total investments amount and the like, which are not limited in the application.
The influence of the individual features may have a positive or negative influence.
On the basis of the foregoing embodiment, optionally, before obtaining the prediction result corresponding to the prediction sample through the financial product marketing model, the method further includes:
determining target characteristics in all characteristics of the financial product marketing model;
the user can select one or more important characteristics as target characteristics from all the characteristics of the financial product marketing model, wherein the specific number is based on actual operation requirements, for example, 100 characteristics are selected as the total characteristics, and 10 characteristics are selected as the target characteristics.
Generating, by a machine learning model interpretation tool, at least one influence trend graph that matches the financial product marketing model, comprising:
and respectively generating influence trend graphs corresponding to the target characteristics of the financial product marketing model through a machine learning model interpretation tool.
Specifically, a specific feature is selected by the SHAP tool to generate an influence trend graph of the corresponding feature.
On the basis of the foregoing embodiment, optionally, the influence trend graph is used to represent a trend of the influence of the target feature on the prediction result as the feature value of the target feature changes.
The target feature may have a positive influence or a negative influence on the prediction result as the feature value of the target feature changes, and in a certain interval range, it can be observed that the probability that the feature value has a negative influence or a positive influence in the interval is high.
In the actual operation process of the model, after a new prediction sample is input into the financial product marketing model, at least one characteristic interpretation chart and at least one characteristic influence chart matched with the prediction sample are generated according to the SHAP tool.
On the basis of the foregoing embodiment, optionally, the feature interpretation graph is used to represent a trend that influence of the target feature on the prediction result changes with a feature value of the target feature in a process of predicting the prediction sample by the financial product marketing model to obtain the prediction result.
On the basis of the foregoing embodiment, optionally, the feature impact diagram is used to represent the influence contribution degree of each feature to the prediction result or the influence contribution degree of the candidate feature to the prediction result in the process of predicting the prediction result by the financial product marketing model to obtain the prediction result;
the influence contribution degrees comprise positive influence contribution degrees and negative influence contribution degrees, and the candidate features are a preset number of features selected from the features according to the absolute value ranking of the influence contribution degrees.
The influence contribution degree has positive or negative values, the positive value represents the positive influence contribution degree, the negative value represents the negative influence contribution degree, therefore, the absolute value represents the influence, the preset number of characteristics with the size of the absolute value ranked at the top is selected, the preset number can be determined according to the actual situation, and the optional preset number is smaller than the number of all the characteristics.
And S130, generating a model interpretation report according to the influence global graph, the influence trend graph, the feature interpretation graph and the feature influence graph, so that the user can carry out financial product marketing according to the prediction result and the model interpretation report.
The model interpretation report comprises an influence global graph, an influence trend graph, a characteristic interpretation graph and a characteristic influence graph and quantitative analysis interpretation of corresponding images.
Illustratively, the report is generated by performing a quantitative analysis on the global graph of influence. Referring to fig. 2, by normalizing the SHAP value of the feature value IN the interval of-1.25 to 0.75, the darker the color represents a greater probability of negative influence, and the grayer the color represents a greater positive influence, it can be observed that the maximum income transaction amount (feature IN _ MAX) of the sample exhibits a uniform distribution of two grays and blacks IN the interval, which indicates that the recent income amount of the sample has a small influence on the prediction result, and after the report is generated, the user can choose to reduce the weight of the feature IN the model training or choose to delete the feature according to the description of the feature IN the report; it can be observed from the figure that the recent expenditure amount (feature OUT _ MAX) of the sample is substantially black in the range from-0.25 to 0, and is substantially gray in the range from 0 to 0.25, and the user can choose to market the sample with a large recent expenditure amount if the positive influence on the marketing success probability of the sample is larger when the recent expenditure amount of the sample is larger. In addition, it can be observed from the figure that the higher the invested total capital sum (characteristic TOT _ IVST _ BAL) of the customer is, the greater the negative impact on the marketing success probability is, and the user can avoid increasing marketing activities on the sample with high invested total capital sum at this stage.
Illustratively, the report is generated by performing quantitative analysis on a feature interpretation graph of a feature monthly daily average asset management scale (feature MAVER _ AUM _ BAL).
As shown in fig. 3, black represents a negative effect, and gray represents a positive effect, and it can be known that the smaller the asset management scale, the more positive the effect on the marketing success, and the more suitable the marketing of the financial product; on the contrary, the larger the asset management scale is, the more negative the impact is, and the less financial products are recommended to be marketed.
By way of example, the report may also be generated by performing a quantitative analysis on a characteristic impact diagram of a sample.
As in fig. 4, black represents a positive effect, gray represents a negative effect, and a larger color area indicates a larger influence. It can be observed that, the sample's LAST three months electronic bank login number (signature EBANK _ LAST _3months _logic _) and maximum inflowing fund transaction amount (signature IN _ MAX) have the greatest positive contribution to its marketing success probability, the current fund management size (DAY _ AUM _ BAL) and the total invested fund (TOT _ IVST _ BAL) have the greatest negative impact on their probability of marketing success.
The complete influence of the sample features can be further expanded as shown in fig. 5, where black represents a positive influence and gray represents a negative influence, as can be more clearly observed in fig. 5, the sample's electronic bank login number (characteristic EBANK _ LAST _3months _login _num) in the LAST three months is at a high level, possibly reflecting the sample's high capital operation requirements, which is beneficial for marketing financial products to it. The user can purposely spread marketing around the information, thereby improving the success probability.
According to the technical scheme of the embodiment of the invention, the model interpretation report is generated through the influence global graph, the influence trend graph, the characteristic interpretation graph and the characteristic influence force graph, the interpretative information is visual and easy to read, the post audit is convenient to deal with, the interpretable information can be directly provided for the characteristics no matter what model is used, the interpretation effect is good, the tuning cost is low, and the automation degree is high.
Example two
Fig. 6 is a flowchart of a model interpretation method according to a second embodiment of the present invention, which is optimized based on the second embodiment. As shown in fig. 6, the method of this embodiment specifically includes the following steps:
and S210, obtaining a prediction result corresponding to the prediction sample through a financial product marketing model.
And S220, generating an influence global graph and at least one influence trend graph matched with the financial product marketing model and at least one characteristic interpretation graph and at least one characteristic influence graph matched with the prediction sample through a machine learning model interpretation tool.
And S230, generating a model interpretation report according to the influence global graph, the influence trend graph, the feature interpretation graph and the feature influence graph, so that the user can carry out financial product marketing according to the prediction result and the model interpretation report.
By way of example, the impact trend graph is analyzed quantitatively to generate a report. As shown in fig. 7, black represents a positive influence, gray represents a negative influence, the probability of the positive influence is relatively high in the interval 60 to 80, and the probability of the negative influence is relatively high in the interval 160 to 180.
S240, determining the characteristics to be adjusted according to at least one model interpretation report; and adjusting parameters of the characteristics to be adjusted in the financial product marketing model.
Referring to fig. 7, the probability of the positive influence between 60 and 80 is approximately between 80% and 90%, and the probability of the negative influence is 10% to 20%, but if it is found in the subsequent application process that the probability of the negative influence is relatively large in the interval between 60 and 80 in most sample model interpretation reports, it may be stated that the influence trend graph of the monthly-day-average asset management scale (feature MAVER _ AUM _ BAL) obtained by the model is low in accuracy, and the user may adjust the monthly-day-average asset management scale parameter in the model accordingly. There may also be some feature discovered during the application of the model, the change in the feature value has little effect on the model, and the weight of the feature may be reduced appropriately or the feature may be deleted. If the influence of a certain feature on the prediction result in the actual prediction process is found to be larger than the influence reflected in the influence trend graph, the user can increase the weight of the feature.
According to the method and the device, the characteristics to be adjusted are determined according to at least one model interpretation report, parameters of the characteristics to be adjusted in the financial product marketing model are adjusted, accuracy and effectiveness of model prediction results are improved, and user experience is improved.
EXAMPLE III
Fig. 8 is a schematic structural diagram of a model interpretation apparatus according to a third embodiment of the present invention. As shown in fig. 8, the apparatus includes:
the prediction result obtaining module 810 is configured to obtain a prediction result corresponding to the prediction sample through a financial product marketing model;
a model interpretation module 820 for generating an influence global graph and at least one influence trend graph matched with the financial product marketing model, and at least one feature interpretation graph and at least one feature influence graph matched with the prediction sample through a machine learning model interpretation tool;
and the model interpretation report generation module 830 is used for generating a model interpretation report according to the influence global graph, the influence trend graph, the characteristic interpretation graph and the characteristic influence graph, so that the user can carry out financial product marketing according to the prediction result and the model interpretation report.
On the basis of the above embodiment, optionally, the influence global graph is used for representing each feature of the financial product marketing model, and the influence on the prediction result changes along with the change of the relative value of the feature value.
On the basis of the foregoing embodiment, optionally, the apparatus further includes:
the target characteristic determining module is used for determining target characteristics in all characteristics of the financial product marketing model;
a model interpretation module 820 comprising:
and the influence trend graph generating unit is used for respectively generating the influence trend graphs corresponding to the target characteristics of the financial product marketing model through a machine learning model interpretation tool.
On the basis of the foregoing embodiment, optionally, the influence trend graph is used to represent a trend of the influence of the target feature on the prediction result as the feature value of the target feature changes.
On the basis of the foregoing embodiment, optionally, the feature interpretation graph is used to represent a trend that influence of the target feature on the prediction result changes with a feature value of the target feature in a process of predicting the prediction sample by the financial product marketing model to obtain the prediction result.
On the basis of the foregoing embodiment, optionally, the feature impact diagram is used to represent the influence contribution degree of each feature to the prediction result or the influence contribution degree of the candidate feature to the prediction result in the process of predicting the prediction result by the financial product marketing model to obtain the prediction result;
the influence contribution degrees comprise positive influence contribution degrees and negative influence contribution degrees, and the candidate features are a preset number of features selected from the features according to the absolute value ranking of the influence contribution degrees.
On the basis of the foregoing embodiment, optionally, the apparatus further includes:
the to-be-adjusted feature determining module is used for determining the to-be-adjusted feature according to the at least one model interpretation report;
and the model adjusting module is used for adjusting the parameters of the characteristics to be adjusted in the financial product marketing model.
The model interpretation device provided by the embodiment of the invention can execute the model interpretation method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the execution method.
Example four
FIG. 9 illustrates a schematic diagram of an electronic device 10 that may be used to implement embodiments of the present invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 9, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, and the like, wherein the memory stores a computer program executable by the at least one processor, and the processor 11 can perform various suitable actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from a storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data necessary for the operation of the electronic apparatus 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.
A number of components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, or the like; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.
The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (cpu), a Graphics Processing Unit (GPU), various dedicated Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, and so forth. The processor 11 performs the various methods and processes described above, such as the model interpretation method.
In some embodiments, the model interpretation method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the model interpretation method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the model interpretation method by any other suitable means (e.g. by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
Computer programs for implementing the methods of the present invention can be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be performed. A computer program can execute entirely on a machine, partly on a machine, as a stand-alone software package partly on a machine and partly on a remote machine or entirely on a remote machine or server.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS service are overcome.
It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present invention may be executed in parallel, sequentially, or in different orders, and are not limited herein as long as the desired results of the technical solution of the present invention can be achieved.
The above-described embodiments should not be construed as limiting the scope of the invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A model interpretation method, comprising:
obtaining a prediction result corresponding to the prediction sample through a financial product marketing model;
generating, by a machine learning model interpretation tool, an influence global graph and at least one influence trend graph matched with the financial product marketing model, and at least one feature interpretation graph and at least one feature influence graph matched with the prediction samples;
and generating a model interpretation report according to the influence global graph, the influence trend graph, the characteristic interpretation graph and the characteristic influence graph so as to enable the user to carry out financial product marketing according to the prediction result and the model interpretation report.
2. The method of claim 1, wherein the global impact map is used to represent the impact of each feature of the financial product marketing model on the predicted outcome as the relative value of the feature changes.
3. The method of claim 1, further comprising, prior to obtaining the predicted outcome corresponding to the predicted sample via a financial product marketing model:
determining target characteristics in all characteristics of the financial product marketing model;
generating, by a machine learning model interpretation tool, at least one influence trend graph that matches the financial product marketing model, comprising:
and respectively generating influence trend graphs corresponding to the target characteristics of the financial product marketing model through a machine learning model interpretation tool.
4. The method of claim 3, wherein the influence trend graph is used for representing the influence of the target feature on the prediction result as the change trend of the feature value of the target feature.
5. The method of claim 1, wherein the feature interpretation graph is used for representing a trend that influence of the target feature on the prediction result changes along with the feature value of the target feature in the process of predicting the prediction sample by the financial product marketing model to obtain the prediction result.
6. The method of claim 1, wherein the feature impact diagram is used for representing the degree of contribution of each feature to the prediction result or the degree of contribution of the candidate features to the prediction result in the process of predicting the prediction result by the financial product marketing model to obtain the prediction result;
the influence contribution degrees comprise positive influence contribution degrees and negative influence contribution degrees, and the candidate features are a preset number of features selected from the features according to the absolute value ranking of the influence contribution degrees.
7. The method of claim 1, after generating the model interpretation report, further comprising:
determining features to be adjusted according to at least one model interpretation report;
and adjusting parameters of the characteristics to be adjusted in the financial product marketing model.
8. A model interpretation apparatus, characterized in that the apparatus comprises:
the prediction result acquisition module is used for acquiring a prediction result corresponding to the prediction sample through the financial product marketing model;
a model interpretation module for generating, by a machine learning model interpretation tool, a global graph of influence and at least one trend graph of influence matching the financial product marketing model, and at least one feature interpretation graph and at least one feature impact graph matching the prediction samples;
and the model interpretation report generation module is used for generating a model interpretation report according to the influence global graph, the influence trend graph, the characteristic interpretation graph and the characteristic influence graph so as to enable a user to carry out financial product marketing according to the prediction result and the model interpretation report.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the model interpretation method according to any of claims 1 to 7 when executing the program.
10. A storage medium storing computer-executable instructions for performing the model interpretation method of any one of claims 1 to 7 when executed by a computer processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211337770.6A CN115759283A (en) | 2022-10-28 | 2022-10-28 | Model interpretation method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211337770.6A CN115759283A (en) | 2022-10-28 | 2022-10-28 | Model interpretation method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115759283A true CN115759283A (en) | 2023-03-07 |
Family
ID=85355906
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211337770.6A Pending CN115759283A (en) | 2022-10-28 | 2022-10-28 | Model interpretation method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115759283A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117710066A (en) * | 2024-02-05 | 2024-03-15 | 厦门傲凡科技股份有限公司 | Financial customer recommendation method and system |
-
2022
- 2022-10-28 CN CN202211337770.6A patent/CN115759283A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117710066A (en) * | 2024-02-05 | 2024-03-15 | 厦门傲凡科技股份有限公司 | Financial customer recommendation method and system |
| CN117710066B (en) * | 2024-02-05 | 2024-05-10 | 厦门傲凡科技股份有限公司 | Financial customer recommendation method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20200202428A1 (en) | Graphical structure model-based credit risk control | |
| CN111340616B (en) | Method, device, equipment and medium for approving online loan | |
| CN112561082A (en) | Method, device, equipment and storage medium for generating model | |
| CN114118287A (en) | Sample generation method, sample generation device, electronic device and storage medium | |
| CN113420792A (en) | Training method of image model, electronic equipment, road side equipment and cloud control platform | |
| CN115759283A (en) | Model interpretation method and device, electronic equipment and storage medium | |
| CN114494814A (en) | Attention-based model training method and device and electronic equipment | |
| CN112580800A (en) | Training method of data trend prediction model and prediction method of data trend | |
| CN117333076A (en) | Evaluation method, device, equipment and medium based on mixed expert model | |
| CN114999665A (en) | Data processing method and device, electronic equipment and storage medium | |
| CN110705654B (en) | Method, apparatus, electronic device, and medium for monitoring assets | |
| Rosolia et al. | Analyzing deep generated financial time series for various asset classes | |
| CN114331379B (en) | Method for outputting task to be handled, model training method and device | |
| CN117474669A (en) | Loan overdue prediction method, device, equipment and storage medium | |
| CN116974621A (en) | Parameter configuration method, device and equipment | |
| CN117591567A (en) | Asset information analysis method and device | |
| CN115600129A (en) | Information identification method and device, electronic equipment and storage medium | |
| CN114490817A (en) | Traffic classification method, device, electronic equipment, medium and product | |
| CN117454161A (en) | Network model training method, device, equipment and storage medium | |
| CN117035846A (en) | Information prediction method and device and related equipment | |
| CN117892877A (en) | Mobile phone banking user behavior prediction method, device, equipment and medium | |
| CN115239501A (en) | Method and device for determining transaction deposit and electronic equipment | |
| CN117593113A (en) | Credit card account risk assessment method, apparatus, device and storage medium | |
| CN115034893A (en) | Deep learning model training method, risk assessment method and device | |
| CN117651167A (en) | Resource recommendation method, device, equipment and storage medium |
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 |