Method and apparatus for constructing a model

Abstract

An embodiment of this specification provides a method for constructing a model. The method includes: First, receive the defining data input by the user based on the interaction interface for constructing a first model. The first model is used to score each of a plurality of samples. Specifically, it includes at least: receiving at least one preset condition, where each preset condition is a condition that the score result required for the first model to score at least one of the plurality of samples satisfies, and obtaining at least one sample feature selected from a plurality of sample features for use as the input feature of the first model. Then, construct the first model.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of computer technology, and more specifically, to a method and apparatus for building a model. Background Art

[0002] Evaluating the risk of samples or assessing the quality of samples by building models is the daily work of many data analysts. Building a model needs to be achieved through a modeling tool. However, existing modeling tools are usually too professional, which makes data analysts need to spend a lot of time and energy learning relevant background knowledge. At the same time, the advantages of data analysts in business cannot be brought into play.

[0003] Therefore, there is an urgent need for an improved modeling solution that can allow data analysts to concentrate on business analysis, enable data analysts to build and modify models according to their own needs at a relatively low learning cost, so as to meet the modeling needs of their daily work. Summary of the Invention

[0004] This specification describes a method and apparatus for building a model, which can enable data analysts to quickly get started with extremely low learning costs. At the same time, during the entire modeling process, data analysts do not need to understand extra knowledge outside of business, and can concentrate all their energy on business analysis, giving full play to their advantages in business experience, and building and modifying models according to their own needs to meet the modeling needs of their daily work.

[0005] According to a first aspect, there is provided a method for building a model, the method comprising: receiving, by an interaction interface, user-entered qualification data for building a first model, the first model being used to score each of a plurality of samples; the receiving of the qualification data for building the first model at least includes: receiving at least one preset condition entered, where each preset condition is a condition that the score result of the first model for at least one of the plurality of samples is required to satisfy; obtaining at least one sample feature selected from a plurality of sample features for use as an input feature of the first model; and building the first model.

[0006] According to a second aspect, there is provided an apparatus for constructing a model, the apparatus comprising: a receiving unit configured to receive qualification data input by a user based on an interaction interface for constructing a first model, the first model being used to score each of a plurality of samples, the receiving unit at least including: a first receiving module configured to receive at least one preset condition input, where each preset condition is a condition that the score result of the first model for at least one of the plurality of samples is required to satisfy; a second receiving module configured to obtain at least one sample feature selected from a plurality of sample features for use as an input feature of the first model. A constructing unit configured to construct the first model.

[0007] According to a third aspect, there is provided a computer-readable storage medium having a computer program stored thereon, which when executed on a computer causes the computer to execute the method of the first aspect.

[0008] According to a fourth aspect, there is provided a computing device including a memory and a processor, characterized in that an executable code is stored in the memory, and when the processor executes the executable code, the method of the first aspect is implemented.

[0009] By using the method and apparatus for constructing a model provided in the embodiments of the present specification, data analysts can quickly get started with extremely low learning costs. At the same time, during the entire modeling process, data analysts do not need to understand additional knowledge outside of the business, and can concentrate all their energy on business analysis, giving full play to their advantages in business experience, and establishing and modifying models according to their own needs to meet the modeling needs of their daily work. Description of the Drawings

[0010] In order to more clearly illustrate the technical solutions of the multiple embodiments disclosed in this specification, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only the multiple embodiments disclosed in this specification. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0011] Figure 1 A flowchart showing a method for constructing a model according to an embodiment;

[0012] Figure 2A A schematic diagram of an interaction interface disclosed in an embodiment of the present specification;

[0013] Figure 2B A second schematic diagram of an interaction interface disclosed in an embodiment of the present specification;

[0014] Figure 2C A third schematic diagram of an interaction interface disclosed in an embodiment of the present specification;

[0015] Figure 3A The fourth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0016] Figure 3B The fifth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0017] Figure 4A The sixth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0018] Figure 4B The seventh schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0019] Figure 5A The eighth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0020] Figure 5B The ninth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0021] Figure 6A The tenth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0022] Figure 6B The eleventh schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0023] Figure 6C The twelfth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0024] Figure 7 The thirteenth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0025] Figure 8 The fourteenth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0026] Figure 9 The fifteenth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0027] Figure 10 The sixteenth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0028] Figure 11 The seventeenth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0029] Figure 12 The eighteenth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0030] Figure 13 The nineteenth schematic diagram of the interaction interface disclosed in the embodiments of this specification;

[0031] Figure 14 It is the twentieth schematic diagram of the interactive interface disclosed in the embodiments of this specification;

[0032] Figure 15 It is the twenty - first schematic diagram of the interactive interface disclosed in the embodiments of this specification;

[0033] Figure 16 It shows the structural diagram of the device for building a model according to an embodiment. Detailed implementation manners

[0034] Next, in combination with the accompanying drawings, multiple embodiments disclosed in this specification will be described.

[0035] As mentioned above, currently, the modeling tools used by data analysts, such as stock traders, etc., are too professional, resulting in too high learning costs for data analysts and unable to give full play to their own advantages in business. Based on this, the embodiments of this specification provide a method for building a model, which can enable data analysts to quickly get started with extremely low learning costs. At the same time, during the entire modeling process, data analysts do not need to understand extra knowledge outside of business and can concentrate all their energy on business analysis, such as stock analysis, etc., giving full play to their advantages in business experience, and establishing and modifying models according to their own needs to meet the modeling needs of their daily work. Next, in combination with specific embodiments, the implementation steps of the method will be described.

[0036] Specifically, Figure 1 It shows the flowchart of the method for building a model according to an embodiment. The execution subject of the method can be a device with processing capabilities: a system or a device or a processing platform, for example, a client, etc. As Figure 1 shown, the method process includes the following steps: Step S110, receiving a selection instruction for a first sample set sent by the user based on the interactive interface, where the first sample set includes multiple samples. Step S120, displaying the sample data corresponding to each sample in the multiple samples, where the sample data includes the feature values of each sample corresponding to each sample feature among multiple sample features. Step S130, receiving the defined data input by the user for building a first model, where the first model is used to score each sample, and step S130 at least includes: Step S131, receiving at least one preset condition, where each preset condition requires the score result of the first model for scoring at least one sample in the multiple samples to meet the condition; Step S132, obtaining at least one sample feature selected from multiple sample features for use as the input feature of the first model. Step S140, building the first model.

[0037] It should be noted that, through the above steps S110 and S120, the viewing and previewing of the target sample data can be realized. Through step S130, the data analyst can input the model-limited data, which at least includes the setting of preset conditions and model input features. The data analyst only needs to rely on the sample information at hand and give full play to their own business experience and business analysis capabilities to complete the setting of the limited data. Through step S140, a first model can be constructed based on the limited data, including model input features and corresponding model parameters. In addition, steps S110 and S120 are optional, and users can view the sample data through other channels (such as excel sheets or paper documents, etc.).

[0038] The above steps are specifically as follows:

[0039] First, in step S110, a selection instruction for the first sample set is received from the user based on the interaction interface, and the first sample set includes multiple samples. And, in step S120, the sample data corresponding to each sample in the multiple samples is displayed, and the sample data includes the feature values of each sample corresponding to each sample feature among multiple sample features.

[0040] Through steps S110 and S120, the selection of the first sample set and the preview of the corresponding sample data can be realized.

[0041] In one embodiment, the fields involved in the first sample set can be various. For example, stock evaluation, employee performance assessment, food nutritional value analysis, etc. In a specific embodiment, the first sample set belongs to the field of stock evaluation, and the multiple sample features or multiple evaluation indicators corresponding to each sample can include earnings per share, operating income, net profit, cash flow per share, gross profit margin, etc. In another specific embodiment, the first sample set is related to employee business assessment, and the multiple sample features or multiple assessment indicators corresponding to each sample can include attendance rate, workload, etc.

[0042] In one embodiment, a sample set dropdown box including multiple alternative sample sets can be designed in the interaction interface (such as the client interface of the modeling tool) for the user (including data analysts) to select the first sample set from. Specifically, step S110 can include: receiving the selection instruction for the first sample set from the user based on the sample set dropdown box. In another embodiment, an input box for inputting the sample set name can be designed in the interaction interface.

[0043] In one embodiment, step S120 can include: displaying the sample data corresponding to each sample in the multiple samples in the first sample set in the form of a list.

[0044] In an example, it can be throughFigure 2A In the sample set drop-down box 210 shown in the figure, select the sample set with the table name stock. Based on the selection instruction issued thereby, display the relevant sample data in the form of a list, such as Figure 2A shown. Among them, for the stock with the sample name or stock name of Technology No. 1, the eigenvalue corresponding to the sample feature earnings per share is 0.010929 yuan. In this way, the sample data of the first sample set can be displayed.

[0045] Furthermore, on the one hand, in one embodiment, after step S120, it may further include: in response to a collapse instruction for the sample data, perform a collapsed display on the sample data. In one example, as Figure 2B shown, the collapse instruction can be issued by clicking the collapse list details icon 220. Further, the expand instruction can also be issued by clicking the expand list details icon 230 to view the sample data.

[0046] On the other hand, in one embodiment, after step S120, it may further include: updating and displaying the sample data of multiple samples as data after normalization processing. Specifically, the normalization processing refers to performing normalization processing on multiple eigenvalue corresponding to each sample feature. In a specific embodiment, the interactive interface includes a normalization processing icon and a corresponding option box. Based on this, in response to the confirmation selection instruction for the option box, use the data obtained by performing normalization processing on the multiple eigenvalue corresponding to each sample feature of each sample to update and display the sample data. In one example, as Figure 2C shown, based on the trigger of the option box 240, display the updated sample data, where the eigenvalue columns corresponding to each sample feature are all updated to the corresponding eigenvalue columns after normalization processing. For example, the sum of the multiple eigenvalue corresponding to the "earnings per share (yuan)" column is 1.

[0047] As above, based on step S110 and step S120, the selection and preview of the first sample set can be realized. Next, in step S130, receive the limited data input by the user for constructing the first model, and the first model is used to score each sample.

[0048] Specifically, step S130 includes at least step S131 and step S132. In step S131, at least one preset condition input by the user is received. In one embodiment, the number of input preset conditions can be one or more, where the preset conditions can be set by the data analyst according to the sample information at hand. Each preset condition is a condition that the score result required by the first model for at least one of the multiple samples needs to meet. In a specific embodiment, the sample information can include the sample data shown above, and can also include other information related to the samples. In an example, the sample information can include Figure 2A the index data of the stocks shown in, and can also include industry data related to the stocks, or more specific and detailed data such as the historical trend curves of some of the stocks. In this way, the data analyst can conduct business analysis based on the sample information and set preset conditions in combination with their own experience.

[0049] In a specific embodiment, at least one of the set preset conditions may include a first preset condition of the final score type. The first preset condition includes at least one of the multiple samples, a first operator, and a preset score. The first operator is one of greater than, less than, and equal to. The first preset condition is used to define the relative magnitude between the score obtained by using the first model to score each of the at least one sample and the preset score. In an example, at least one of the samples included in the first preset condition is Figure 2A the four stocks numbered Education 1 to Education 4 shown in, the first operator is greater than (>), and the predetermined score is 80 points.

[0050] Further, a condition type drop-down box and a condition template corresponding to the final score type can be designed in the interaction interface for the user to conveniently and quickly input preset conditions. In an example, the condition type drop-down box in the interaction interface includes multiple alternative types, including the above-mentioned final score type. Based on this, receiving the input first preset condition may include: First, receiving a selection instruction for the final score type issued based on the condition type drop-down box; then, displaying a first condition template corresponding to the final score type. The first condition template includes a sample drop-down box, an operator drop-down box, and an operation value input box. The sample drop-down box includes the multiple samples for guiding the user to select the at least one sample therefrom. The operator drop-down box includes multiple alternative operators for guiding the user to select the first operator therefrom. The operation value input box is used to guide the user to input the preset score; then, receiving the first predetermined condition input by the user based on the first condition template and incorporating it into the at least one preset condition. In a specific example, such as Figure 3AAs shown, first, receive the selection instruction for the final score category in the condition type drop-down box 310; then, display the corresponding first condition template 320; then, receive the first predetermined condition input by the user based on the first condition template 320, where the first predetermined condition is that the final model scores of Technology No. 1 and Technology No. 2 are greater than 60 points.

[0051] Further, the first preset condition further includes a weight for indicating the importance degree of the condition, and the first condition template further includes a weight input box and / or a weight drag bar for guiding the user to input the weight. In one example, as Figure 3B shown, the user can set the weight of the first preset condition through the weight drag bar 330. In this way, the first preset condition set by the user can be received.

[0052] In another specific embodiment, at least one of the set preset conditions may include a second preset condition of the relative ranking type, the second preset condition includes a first sample subset, a second sample subset, and a second operator, the samples in the first sample subset and the second sample subset together form the at least one sample, the second operator is one of greater than, less than, and equal to, and the second preset condition is used to define the relative magnitude between the scores obtained by scoring each sample in the first sample subset using the first model and the scores obtained by scoring each sample in the second sample subset. In one example, the samples in the first sample subset include Figure 2A Entertainment No. 1 and Entertainment No. 2 shown in, and the samples in the second sample subset include Figure 2A Education No. 1 shown in, and the second operator is less than (<), and the second preset condition formed thereby means that the score of Entertainment No. 1 and Entertainment No. 2 is less than the score of Education No. 1.

[0053] Furthermore, a condition type dropdown box and a condition template corresponding to the relative ranking class can be designed in the interaction interface. In one example, the condition type dropdown box in the interaction interface includes multiple alternative types, including the above-mentioned relative ranking class. Based on this, receiving the input second preset condition can include: First, receiving a selection instruction for the relative ranking class issued based on the condition type dropdown box; then, displaying the second condition template corresponding to the relative ranking class, where the second condition template includes a first sample dropdown box, a second sample dropdown box, and an operator dropdown box. The first sample dropdown box includes the multiple samples and is used to guide the user to select samples therefrom to form the first sample subset. The second sample dropdown box includes the multiple samples and is used to guide the user to select samples therefrom to form the second sample subset. The operator dropdown box includes multiple alternative operators and is used to guide the user to select the second operator therefrom; then, receiving the second predetermined condition input by the user based on the second condition template and classifying it into the at least one preset condition. In a specific example, as Figure 4A shown, first, receive a selection instruction for the relative ranking class in the condition type dropdown box 410; then, display the corresponding second condition template 420; then, receive the second predetermined condition input by the user based on the second condition template 420, where the second predetermined condition is that the scores of Entertainment No. 1 and Entertainment No. 2 are less than the score of Education No. 1.

[0054] Furthermore, the second preset condition also includes a weight indicating the importance degree of the condition. The second condition template also includes a weight input box and / or a weight drag bar for guiding the user to input the weight. In one example, as Figure 4B shown, the user can set the weight of the second preset condition through the weight drag bar 430. In this way, the second preset condition set by the user can be received.

[0055] In addition, in one example, each of the at least one preset conditions above can include a corresponding weight for guiding the construction of the first model. For example, the at least one preset condition includes a first preset condition and a second preset condition. In the process of constructing the first model, if these two preset conditions cannot be satisfied simultaneously, the preset condition with a greater weight can be preferentially satisfied.

[0056] On the other hand, in one embodiment, the preset condition can include a default condition preset by the system developer. For example, the valid range of the sample score, such as -10^8 to 10^8.

[0057] The above can receive at least one preset condition input by the user in step S131. Then in step S132, at least one sample feature selected from multiple sample features is obtained for use as the input feature of the first model.

[0058] In one embodiment, a plurality of feature icons corresponding to a plurality of sample features and a filtering box corresponding to each of the feature icons are designed in the interaction interface. In this way, the user can confirm the selection and deselection of the sample features through the filtering box. Specifically, step S132 may include: receiving a confirmation selection instruction for at least one filtering box corresponding to the at least one sample feature. In one example, Figure 5A as shown in, a plurality of filtering boxes corresponding to a plurality of feature icons are shown, and by clicking on some of the filtering boxes, corresponding partial sample features can be selected, including earnings per share (yuan), operating income (ten thousand yuan), year-on-year growth rate of operating income (%), etc.

[0059] In addition, in one embodiment, in order to assist the user in selecting sample features, an acquisition icon for the recommendation degree of the sample features may also be designed in the interaction interface. Based on this, after step S131 and after step S132, the following may also be included: receiving a trigger instruction for the acquisition icon; displaying the recommendation degree of each of the plurality of sample features to assist the user in selecting the at least one sample feature. Further, in a specific embodiment, the trigger instruction may be a click instruction or a voice control instruction, etc. On the other hand, regarding the determination of the recommendation degree, for example, assuming that the received predetermined conditions include that the score of sample 1 is greater than the score of sample 2, and the score of sample 3 is greater than the score of sample 4, and the above-mentioned plurality of sample features include feature a and feature b. At this time, if for feature a, the feature value of sample 1 is greater than the feature value of sample 2, and the feature value of sample 3 is greater than the feature value of sample 4, the feature recommendation degree of feature a can be determined to be 100%. At the same time, if for feature b, the feature value of sample 1 is less than the feature value of sample 2, and the feature value of sample 3 is greater than the feature value of sample 4, the feature recommendation degree of feature a can be determined to be 50%.

[0060] In one example, as Figure 5B shown, in response to a trigger instruction for the acquisition icon 510, a plurality of feature recommendation degrees corresponding to a plurality of sample features are displayed. For example, the recommendation degree of the sample feature earnings per share is 100%.

[0061] As above, through step S131 and step S132, at least one preset condition in the defined data and at least one sample feature as the input feature of the model can be received.

[0062] In addition, in one embodiment, the model type of the first model is a linear model. Correspondingly, the first model includes at least one model parameter corresponding to at least one sample feature. Based on this, the received qualification data may further include at least one parameter value interval for the at least one sample feature. Specifically, in step S130, it may further include: receiving at least one parameter value interval set for the at least one sample feature, where the at least one parameter value interval is used to define the value range of the at least one model parameter.

[0063] Further, in a specific embodiment, the interaction interface is designed with a plurality of adjustment icons for uniformly setting the upper and lower limits of the plurality of parameter value intervals of the plurality of sample features, including an upper limit uniformly increasing icon, an upper limit uniformly decreasing icon, a lower limit uniformly increasing icon, and a lower limit uniformly decreasing icon. Based on this, receiving the at least one parameter value interval set above may include: in response to a trigger instruction for any first adjustment icon among the plurality of adjustment icons, uniformly adjusting the plurality of parameter value intervals corresponding to the first adjustment icon. The plurality of parameter value intervals include the at least one parameter value interval, and obviously, the adjustment of at least one of the parameter value intervals can be achieved. In addition, each trigger of the icon can be adjusted at a predetermined interval value. In one example, as Figure 6A shown, the interaction interface includes a lower limit uniformly increasing icon 611, a lower limit uniformly decreasing icon 612, an upper limit uniformly increasing icon 621, and an upper limit uniformly decreasing icon 622. Further, in response to a click instruction for the lower limit uniformly increasing icon 611 among them, the lower limits of the plurality of parameter value intervals are all changed from -10 to -5 at a predetermined interval value of 5. In this way, the setting of at least one parameter value interval can be completed.

[0064] In another embodiment, the received qualification data may further include a model parameter set for a part of the at least one sample feature. Specifically, the at least one sample feature includes a first sample feature, and the at least one model parameter includes a first model parameter corresponding to the first sample feature. Based on this, in step S130, it may further include: receiving the first model parameter set for the first sample feature.

[0065] Further, in a specific embodiment, the interaction interface is designed with a plurality of parameter value input boxes and / or a plurality of parameter value sliders corresponding to the plurality of sample features, including a first parameter value input box and / or a first parameter value slider corresponding to the first sample feature. Based on this, receiving the first model parameter above may include: receiving the first model parameter set based on the first parameter value input box and / or the first parameter value slider. In one example, Figure 6BThe shown interactive interface includes multiple parameter value input boxes and multiple parameter value sliders, specifically including a parameter value input box 630 and a parameter value slider 640 corresponding to the sample feature earnings per share.

[0066] In a more specific embodiment, to prevent the above-mentioned first parameter value from being accidentally edited, it can also be locked. Specifically, the interactive interface further includes multiple lock icons corresponding to the multiple parameter value input boxes, including a first lock icon corresponding to the first parameter value input box, and the first lock icon is in an unlocked state; after receiving the first model parameter set based on the first parameter value input box, the method further includes: receiving a trigger instruction for the first lock icon; updating and displaying the unlocked first lock icon as a locked state, and displaying the first parameter value input box displaying the first model parameter as an uneditable state. In an example, Figure 6C The shown interactive interface includes a first lock icon 650 in an unlocked state, and in response to a click instruction on it, it is updated and displayed as a locked state. It can be understood that clicking again can unlock it. In this way, the locking of the first model parameter can be achieved.

[0067] In yet another embodiment, the received limiting data may further include a model type, such as a linear model or a non-linear model. In an example, such as Figure 7 shown, the interactive interface includes a filtering box 710 corresponding to a linear model icon and a filtering box 720 corresponding to a non-linear model icon. By clicking on the filtering box 720, the deselection of the filtering box 710 and the confirmation selection of the filtering box 720 can be triggered, thereby setting the model type to a non-linear model. It should be noted that the model type may not require the user to make a selection, but the background automatically realizes the filtering of the model type.

[0068] As described above, in step S130, the limiting data input by the user for constructing the first model can be received. The limiting data may include at least one predetermined condition and at least one sample feature used as a model input feature, and may also include a set model type, at least one parameter value interval set for at least one sample feature, and model parameters set for some of the at least one sample features.

[0069] Next, in step S140, the first model is constructed.

[0070] In one embodiment, after receiving the user input of the limiting data, the construction of the first model can be automatically triggered. In another embodiment, the first model can be constructed in response to a construction instruction for the first model. In yet another embodiment, the first model can be constructed in response to a display instruction for the first model for display.

[0071] It should be understood that the model includes input features and model parameters. Accordingly, a first model is constructed, including determining the input features and model parameters of the first model. In one embodiment, at least one sample feature obtained in the foregoing step S132 may be used as the input feature of the first model. Further, the quadratic programming method can be used to determine the model parameters based on the limited data. It should be noted that the quadratic programming method is a special type of mathematical programming problem in nonlinear programming, which is used to find the optimal solution and has applications in many aspects, such as the solution of constrained least squares problems, the application of sequential quadratic programming in nonlinear optimization problems, etc. When the quadratic programming problem has only equality constraints, the quadratic programming can be solved by linear equations. Otherwise, common quadratic programming solution methods include: interior point method, active set, and conjugate gradient method, etc.

[0072] In a specific embodiment, it is assumed that the limited data obtained includes: the set model type is a linear model; two selected sample features, denoted here by x 1 and x 2 ; the preset conditions are that the score of sample A is greater than 80, the score of sample B is equal to 60, and the score of sample C is greater than the score of sample D, which are respectively expressed by the following formulas: y A > 80, y B = 60, y C > y D ; the value ranges respectively set for the model parameters corresponding to the two sample features are

[0073] Based on this, in order to solve the optimal combination of model parameters, the following constraint conditions can be constructed:

[0074]

[0075]

[0076]

[0077]

[0078] In formula (1), and and and and are respectively the scores of sample A, sample B, sample C, and sample D corresponding to the features x 1 and x 2The eigenvalues. Thus, based on formula (1), the model parameters k 1 , k 2 and the optimal combination of b can be solved.

[0079] As described above, based on the obtained limited data, the first model can be constructed.

[0080] It should be noted that in one embodiment, after step S140, it may further include: receiving a display instruction for the first model, and displaying the constructed first model, where the first model includes the at least one sample feature and model parameters.

[0081] In a specific embodiment, the display instruction may be a click instruction or a voice control instruction, etc.

[0082] In one example, in response to Figure 8 a click instruction on the run icon 810 in the shown interaction interface, the first model 820 is displayed. The first model 820 is a linear model, which includes at least one sample feature: earnings per share (yuan), operating income (ten thousand yuan), net profit (ten thousand yuan), net profit year-on-year (%), return on net assets (%), gross profit margin (%), operating income year-on-year (%), and the corresponding at least one model parameter: 5.72, 5.07, 4.26, -6.03,, 2.72, 0.96, -0.27. Thus, the display of the first model can be realized.

[0083] In addition, in one embodiment, as described above, the interaction interface may include multiple parameter value input boxes corresponding to multiple sample features, including at least one parameter value input box corresponding to at least one sample feature. Based on this, after step S140, it may further include: displaying the at least one model parameter in the at least one parameter value input box. It can be understood that the other input boxes in the multiple parameter value input boxes except the at least one parameter value input box may display numerical values or may be displayed as 0. In one example, Figure 9 the shown interaction interface includes multiple parameter value input boxes, where multiple model parameters are correspondingly displayed. For example, the model parameter 5.72 corresponding to earnings per share (yuan) is displayed in the parameter value input box 910.

[0084] Further, at least one of the above parameter value input boxes includes a first parameter value input box, and the at least one model parameter includes a first model parameter. Thus, the modification of the first model parameter made by the user in the first parameter value input box is received; again in response to the display instruction for the first model, the first model rebuilt based on the modified first model parameter is displayed. In a specific embodiment, as described above, the interaction interface may further include a plurality of lock icons corresponding to a plurality of parameter value input boxes, wherein the lock icon can also be used by the user to lock the model parameters set by the user in the corresponding parameter value input box, so that when the first model is rebuilt subsequently, the locked model parameters always remain unchanged. On the contrary, when the user modifies the corresponding model parameter in the parameter value input box but does not lock it through the lock icon, the set value therein will not be directly used as the model parameter when building the model subsequently, that is, this set value will not be used as the limiting data for building the first model. Therefore, the model parameter in the corresponding parameter value input box will be displayed as it changes with the rebuilding of the first model. In an example, as Figure 10 shown, the first model parameter in the first parameter value input box 1010 is modified from 0.96 to 5, and then the lock icon 1020 is triggered to lock it. Clicking the run icon again can obtain the rebuilt first model. For example, the model parameter of the sample feature earnings per share (yuan) changes from 5.72 to 6.23, while the first model parameter is locked at 5. In this way, the model parameters can be displayed in the parameter value input box to facilitate the user to further modify the model parameters, such as modifying the value range, or setting the parameter values in combination with the lock icon.

[0085] On the other hand, as described above, the first model is used to score each of the multiple samples. In one embodiment, after step S140, it may further include: displaying a plurality of original scores obtained by scoring the multiple samples using the first model. In an example, Figure 11 shows the scores of multiple samples. Further, the interaction solution provided in the embodiments of this specification can also support the adjustment of the overall distribution range of the original scores. For ease of understanding, first take an example. For example, the original range of multiple original scores is [-2.36, 108.5]. If the user hopes that the scores can be within [0, 100], the multiple original scores can be adjusted accordingly based on [0, 100] so that they are distributed within [0, 100]. More specifically, by first uniformly increasing the multiple original scores by 2.36 points (offset value), and then uniformly multiplying by 0.902 (gain coefficient), the multiple original scores can be uniformly adjusted to scores within [0, 100].

[0086] Specifically, after step S140, it may further include: displaying the original intervals corresponding to the multiple original scores. In a specific embodiment, the original mean and original variance of the multiple original scores may also be displayed simultaneously. Further, after displaying the original intervals, it may further receive an adjusted interval set by the user based on the original intervals; and then replace and display the multiple original scores with multiple adjusted scores obtained by adjusting the multiple original scores based on the adjusted intervals. In a specific embodiment, the original mean and original variance are respectively replaced and displayed with the adjusted mean and adjusted variance corresponding to the multiple adjusted scores.

[0087] In one example, as Figure 12 shown, the original interval corresponding to the multiple original scores is [-3.649, 10.064], and the original mean and original variance are 1.353 and 1.889 respectively. After receiving the adjusted interval [0, 100] set by the user, in response to the trigger instruction for the application icon 1210, the multiple original scores, the original mean, and the original variance are replaced and displayed with multiple adjusted scores, the adjusted mean, and the adjusted variance, and further, the gain value 7.293 and the offset value 3.649 are displayed. In this way, the adjustment of the original scores can be completed.

[0088] On the other hand, the multiple scores corresponding to the multiple samples may also be displayed in a more diverse chart form. The multiple scores may be the multiple original scores as described above, or the multiple adjusted scores as described above. Specifically, the multiple scores may be displayed in the form of a histogram and / or a radar chart on the interactive interface. In one embodiment, the interactive interface may include a histogram icon. After step S140, it may further include: in response to the trigger instruction for the histogram icon, displaying a histogram plotted based on the multiple scores; wherein, the histogram includes at least one bar, and the numerical interval of each bar relative to the abscissa represents the corresponding score interval, and the numerical value relative to the ordinate represents the corresponding number of samples. In one example, as Figure 13 shown, in response to the click instruction for the histogram icon 1310, a histogram plotted based on the multiple scores is displayed, wherein the score interval and the number of samples corresponding to the bar 1320 are [25, 30] and 32 respectively.

[0089] Further, it is also possible to adjust the numerical range represented by the abscissa and / or ordinate of the histogram. In a specific embodiment, the histogram includes a plurality of reference grids, and the interval of each reference grid relative to the abscissa corresponds to a first interval fraction; after displaying the histogram drawn based on the plurality of fractions, the method further includes: receiving a second interval fraction obtained by the user adjusting the first interval fraction; using the second interval fraction to update the display of the histogram. In an example, as Figure 13 shown, the interactive interface includes a step value input box 1310 for the step size (interval fraction relative to the abscissa), in which the first interval fraction is displayed as 5. Based on this, the adjusted second interval fraction 10 can be received, and the histogram is updated using the second interval fraction.

[0090] In another embodiment, the sample scores of some samples among a plurality of samples can be displayed in the form of a radar chart. Specifically, the interactive interface includes a radar chart icon. After step S140, it can further include that the method can also include: in response to a trigger instruction for the radar chart icon, displaying a plurality of feature icons and a plurality of screening boxes corresponding to a plurality of sample features, and a radar chart drawn based on the sample features selected from the plurality of sample features. Among them, the plurality of feature icons are arranged in descending order of the absolute values of a plurality of model parameters (it can be understood that the model parameters other than the above at least one model parameter are all 0), the plurality of feature icons include a first number of selected feature icons, and the first number of screening boxes corresponding to the first number of feature icons are in a selected state; displaying a radar chart drawn based on the first number of feature icons, the radar chart including a first number of radial axes. In a specific embodiment, it can be specified that the radar chart includes at most a predetermined number of radial axes, such as 5 or 6, etc. In an example, as Figure 14 shown, in response to a click instruction for the radar chart icon 1410, the corresponding radar chart is displayed, which includes 5 radial axes, corresponding to the sample features earnings per share (yuan), operating income (ten thousand yuan), net profit (ten thousand yuan), net profit year-on-year (%), and return on net assets respectively.

[0091] Furthermore, it is also necessary to select which samples among the multiple samples will have their scores displayed on the radar chart. Specifically, it is possible to receive a second quantity of samples selected by the user from among the multiple samples; and then display a second quantity of closed curves corresponding to the second quantity of samples in the radar chart. In a specific embodiment, to facilitate the user in selecting the second quantity of samples, the multiple samples can first be sorted in order based on the magnitudes of the multiple scores. In a specific embodiment, the first quantity of radial axes includes a first radial axis corresponding to a first sample feature, and the second quantity of closed curves includes a first closed curve corresponding to a first sample, and the value of the first closed curve on the first radial axis is determined by the product of the weight value of the first sample feature and the feature value of the first sample corresponding to the first sample feature. As Figure 14 shown, receive the second quantity of samples selected by the user, with the sample names being Technology No. 1, Education No. 1, and Technology No. 2 respectively. Among them, the coordinate values of Stock Technology No. 2 corresponding to the above 5 radial axes of the sample features are 6.0, 2.779, 3.213, 1.251, and 1.700 respectively. In this way, it is possible to display the sample scores in the form of a radar chart.

[0092] On the other hand, as mentioned above, the first model is constructed based on the limited data input by the user, and the limited data includes at least one preset condition. However, the actually constructed first model usually cannot fully meet all the preset conditions. Therefore, it is possible to display to the user in the interaction interface the satisfaction or compliance degree of the first model with respect to each condition in the preset conditions, or information that can reflect the satisfaction or compliance degree. Further, for either of the two operation contents corresponding to the operator in some preset conditions, there may be more than one sample. Therefore, such a preset condition can also be split into non - further - divisible preset sub - conditions, and the satisfaction degrees of the first model with respect to each preset sub - condition are respectively displayed.

[0093] Specifically, in one embodiment, the interactive interface further includes a preset condition verification icon. After the above step S140, the method may further include: in response to a trigger instruction for the preset condition verification icon, presenting at least one condition difference corresponding to the at least one preset condition, where each condition difference reflects the compliance degree of the first model with respect to the preset condition corresponding to each condition difference. For ease of understanding, first, an example is given to illustrate the determination of the condition difference. For example, a certain preset condition is that the score of sample A is greater than 5, and the scoring result obtained using the first model is that sample A is 10 points or 15 points. In this case, it can be considered that the scoring result meets this preset condition, and the condition difference is 0. Another example is that a certain preset condition is that the score of sample B is greater than the score of sample C. The scoring result obtained using the first model is that sample B is 5 points and the score of sample C is 5.6. In this case, it can be considered that the scoring result does not meet this preset condition, and the condition difference is set to the difference between the two scores, which is 0.6. Alternatively, determine the score differences corresponding to all cases where the preset conditions are not met, and then perform a normalization process, and use the value obtained from the normalization process as the condition difference. Thus, the condition difference is non - negative. Specifically, when the condition difference is 0, it means the satisfaction degree is 100%, that is, the preset condition is fully met. At the same time, the larger the condition difference, the less the scoring result obtained using the first model meets the preset condition. In one example, as Figure 15 shown, in response to a click instruction for the preset condition verification icon 1510, present multiple condition differences corresponding to multiple preset conditions. Among them, the multiple preset conditions are arranged in order of their importance. The condition difference for the preset condition (Education No. 3 > Entertainment No. 1) is 0, indicating that the scoring result of the first model meets this preset condition.

[0094] In this way, it is possible to present to the user in the interactive interface the satisfaction or compliance situation of the first model with respect to each condition in the preset conditions.

[0095] In summary, by using the method for constructing a model disclosed in the embodiments of this specification, data analysts can quickly get started with extremely low learning costs. At the same time, during the entire modeling process, data analysts do not need to understand additional knowledge outside the business, and can concentrate all their energy on business analysis, giving full play to their advantages in business experience, and establishing and modifying models according to their own needs to meet the modeling requirements of their daily work.

[0096] According to an embodiment of another aspect, a device for constructing a model is further provided. Specifically, Figure 16 shows a device for constructing a model according to an embodiment. As Figure 16As shown, the device 1600 includes: a first receiving unit 1610 configured to receive a selection instruction for a first sample set issued by a user based on an interaction interface, where the first sample set includes a plurality of samples. A first display unit 1620 configured to display sample data corresponding to each sample in the plurality of samples, where the sample data includes eigenvalue corresponding to each sample for each of a plurality of sample features. A second receiving unit 1630 configured to receive limitation data input by the user for constructing a first model, where the first model is used to score each sample, and the second receiving unit at least includes: a first receiving module 1631 configured to receive at least one preset condition input, where each preset condition is a condition that the score result required for the first model to score at least one sample in the plurality of samples satisfies; a second receiving module 1632 configured to obtain at least one sample feature selected from the plurality of sample features for use as an input feature of the first model. A construction unit 1640 configured to construct the first model.

[0097] It should be noted that corresponding to the description in the foregoing method, in the device 1600, the first receiving unit 1610 and the first display unit 1620 are optional.

[0098] In one embodiment, the interaction interface includes a sample set dropdown box, and the sample set dropdown box includes a plurality of alternative sample sets, and the plurality of alternative sample sets include the first sample set; the first receiving unit 1610 is specifically configured to: receive a selection instruction for the first sample set issued by the user based on the sample set dropdown box.

[0099] In one embodiment, the device further includes: a collapsed display unit configured to, in response to a collapse instruction for the sample data, perform a collapsed display on the sample data.

[0100] In one embodiment, the interaction interface includes a normalization processing icon and a corresponding option box, and the device further includes: a sample data update display unit configured to, in response to a confirmation selection instruction for the option box, perform an updated display on the sample data by using data obtained by performing normalization processing on a plurality of eigenvalues corresponding to each sample feature for a plurality of samples.

[0101] In one embodiment, the at least one preset condition includes a first preset condition whose condition type is a final score type, the first preset condition includes the at least one sample, a first operator, and a preset score value, the first operator is one of greater than, less than, and equal to, and the first preset condition is used to define the relative magnitude between the score obtained by scoring each sample in the at least one sample by using the first model and the preset score value.

[0102] Further, in a specific embodiment, the interaction interface includes a condition type dropdown box, and the condition type dropdown box includes multiple alternative types, and the multiple alternative types include the final score type;

[0103] The first receiving module 1631 is specifically configured to: receive a selection instruction for the final score type issued based on the condition type dropdown box; display a first condition template corresponding to the final score type, where the first condition template includes a sample dropdown box, an operator dropdown box, and an operation value input box, the sample dropdown box includes the multiple samples, and is used to guide the user to select the at least one sample therefrom, the operator dropdown box includes multiple alternative operators, and is used to guide the user to select the first operator therefrom, and the operation value input box is used to guide the user to input the preset score; receive a first predetermined condition input by the user based on the first condition template, and classify it into the at least one preset condition.

[0104] In a more specific embodiment, the first preset condition further includes a weight indicating the importance degree of the condition, and the first condition template further includes a weight input box, which is used to guide the user to input the weight.

[0105] In an embodiment, the at least one preset condition includes a second preset condition with a condition type of relative ranking type, the second preset condition includes a first sample subset, a second sample subset, and a second operator, the samples in the first sample subset and the second sample subset together form the at least one sample, the second operator is one of greater than, less than, and equal to, and the second preset condition is used to define the relative magnitude between the scores obtained by scoring each sample in the first sample subset using the first model and the scores obtained by scoring each sample in the second sample subset.

[0106] Further, in a specific embodiment, the interaction interface includes a condition type dropdown box, which includes multiple alternative types, and the relative ranking type is included in the multiple alternative types; the first receiving module 1631 is specifically configured to: receive a selection instruction for the relative ranking type issued based on the condition type dropdown box; display a second condition template corresponding to the relative ranking type, where the second condition template includes a first sample dropdown box, a second sample dropdown box, and an operator dropdown box, the first sample dropdown box includes the multiple samples and is used to guide the user to select samples therefrom to form the first sample subset, the second sample dropdown box includes the multiple samples and is used to guide the user to select samples therefrom to form the second sample subset, and the operator dropdown box includes multiple alternative operators and is used to guide the user to select the second operator therefrom; receive the second predetermined condition input by the user based on the second condition template and classify it into the at least one preset condition.

[0107] Further, in a specific embodiment, the second predetermined condition further includes a weight indicating the importance degree of the condition, and the second condition template further includes a weight input box for guiding the user to input the weight.

[0108] In an embodiment, the interaction interface includes multiple feature icons corresponding to the multiple sample features and a filtering box corresponding to each feature icon, and the second receiving module 1632 is specifically configured to: receive a confirmation selection instruction for at least one filtering box corresponding to at least one sample feature.

[0109] In an embodiment, the interaction interface further includes an acquisition icon for the recommendation degree of the sample feature, and the second receiving unit 1630 further includes:

[0110] A third receiving module, configured to receive a trigger instruction for the acquisition icon;

[0111] A display module, configured to display the recommendation degree of each sample feature among the multiple sample features, at least for assisting the user to select the at least one sample feature.

[0112] In an embodiment, the second receiving unit 1630 further includes: a fourth receiving module, configured to receive at least one parameter value interval set for the at least one sample feature, and the at least one parameter value interval is used to limit the value range of the at least one model parameter.

[0113] Further, in a specific embodiment, the interaction interface includes a plurality of adjustment icons for uniformly setting the upper and lower limits of a plurality of parameter value ranges of the plurality of sample features, including an upper limit uniformly increasing icon, an upper limit uniformly decreasing icon, a lower limit uniformly increasing icon, and a lower limit uniformly decreasing icon; the fourth receiving module is specifically configured to: in response to a trigger instruction for any first adjustment icon among the plurality of adjustment icons, uniformly adjust the plurality of parameter value ranges corresponding to the first adjustment icon.

[0114] In one embodiment, the at least one sample feature includes a first sample feature, and the at least one model parameter includes a first model parameter corresponding to the first sample feature; the second receiving unit 1630 further includes: a fifth receiving module, configured to receive the first model parameter set for the first sample feature.

[0115] Further, in a specific embodiment, the interaction interface includes a plurality of parameter value input boxes corresponding to the plurality of sample features, including a first parameter value input box corresponding to the first sample feature; the fifth receiving module is specifically configured to: receive the first model parameter set based on the first parameter value input box.

[0116] In a more specific embodiment, the interaction interface further includes a plurality of locking icons corresponding to the plurality of parameter value input boxes, including a first locking icon corresponding to the first parameter value input box, and the first locking icon is in an unlocked state; the second receiving unit 1630 further includes: a sixth receiving module, configured to receive a trigger instruction for the first locking icon; a locking state update display module, configured to update and display the first locking icon in the unlocked state as a locked state, and display the first parameter value input box displaying the first model parameter as an uneditable state.

[0117] In one embodiment, the second receiving unit 1630 further includes: a seventh receiving module, configured to receive the model type set for the first model, and the model type is a linear model or a non - linear model.

[0118] In one embodiment, the device 1600 further includes a third receiving unit and a second display unit, wherein the third receiving unit is configured to receive a display instruction for the first model, and the second display unit is configured to display the first model, and the first model includes the at least one sample feature and model parameters.

[0119] In one embodiment, the first model is a linear model, which includes at least one model parameter corresponding to the at least one sample feature. The interaction interface includes at least one parameter value input box corresponding to the at least one sample feature. The device further includes: a third display unit configured to display the at least one model parameter in the at least one parameter value input box.

[0120] In a specific embodiment, the at least one parameter value input box includes a first parameter value input box, and the at least one model parameter includes a first model parameter. After displaying the at least one model parameter in the at least one parameter value input box, the device includes: a fourth receiving unit configured to receive the modification of the first model parameter by the user in the first parameter value input box for reconstructing the first model.

[0121] In one embodiment, the device further includes: a fourth display unit configured to display a plurality of original scores obtained by scoring the plurality of samples using the first model.

[0122] In a specific embodiment, the fourth display unit is further configured to display the original intervals corresponding to the plurality of original scores.

[0123] In a more specific embodiment, the device further includes: a fourth receiving unit configured to receive an adjusted interval set by the user based on the original interval; a fifth display unit configured to replace and display the plurality of original scores with a plurality of adjusted scores obtained by adjusting the plurality of original scores based on the adjusted interval.

[0124] Furthermore, in an example, the sixth display unit is further configured to display the gain value and the offset value of the adjusted interval relative to the original interval.

[0125] In an example, the fourth display unit is further configured to display the original mean and the original variance corresponding to the plurality of original scores; the fifth display unit is further configured to replace and display the original mean and the original variance with the adjusted mean and the adjusted variance corresponding to the plurality of adjusted scores respectively.

[0126] In one embodiment, the device further includes: a sixth display unit configured to display the plurality of scores corresponding to the plurality of samples, where the plurality of scores are determined based on a plurality of original scores obtained by scoring the plurality of samples using the first model.

[0127] Further, in a specific embodiment, the device further includes:

[0128] A histogram display unit, configured to display a histogram plotted based on the multiple scores in response to a trigger instruction for the histogram icon; wherein, at least one histogram bar is included in the histogram, and the numerical interval of each histogram bar relative to the abscissa represents the corresponding score interval, and the numerical value relative to the ordinate represents the corresponding number of samples.

[0129] In a more specific embodiment, a plurality of reference grids are included in the histogram, and the interval of each reference grid relative to the abscissa corresponds to a first interval score; the apparatus further includes: a step size receiving unit, configured to receive a second interval score obtained by the user adjusting the first interval score; the histogram display unit is further configured to update and display the histogram by using the second interval score.

[0130] In a specific embodiment, the sixth display unit is specifically configured to perform a ranking display on the multiple samples based on the multiple scores;

[0131] A radar chart icon is included in the interactive interface, and the apparatus further includes a radar chart display unit, configured to: receive a trigger instruction for the radar chart icon; display at least one feature icon and at least one selection box corresponding to the at least one sample feature, the at least one feature icon being arranged in sequence based on the magnitudes of the at least one model parameter, the at least one feature icon including a first number of selected feature icons, and the at least one selection box corresponding to the first number of feature icons being in a selected state; display a radar chart plotted based on the first number of feature icons, the radar chart including a first number of radial axes; receive a second number of samples selected by the user from the multiple samples based on the ranking display; display a second number of closed curves corresponding to the second number of samples in the radar chart; wherein, the first number of radial axes includes a first radial axis corresponding to a first sample feature, and the second number of closed curves includes a first closed curve corresponding to a first sample, and the value of the first closed curve on the first radial axis is determined by the product of the weight value of the first sample feature and the feature value of the first sample corresponding to the first sample feature.

[0132] In an embodiment, a preset condition verification icon is further included in the interactive interface, and the apparatus further includes: a condition difference display unit, configured to display at least one condition difference corresponding to the at least one preset condition in response to a trigger instruction for the preset condition verification icon, wherein each condition difference reflects the compliance degree of the first model with respect to the preset condition corresponding to each condition difference.

[0133] In summary, by using the device for constructing a model disclosed in the embodiments of this specification, data analysts can quickly get started with extremely low learning costs. At the same time, during the entire modeling process, data analysts do not need to understand extra knowledge outside of the business, and can focus all their energy on business analysis, giving full play to their advantages in business experience, and establishing and modifying models according to their own needs to meet the modeling requirements of their daily work.

[0134] As described above, according to an embodiment of still another aspect, there is also provided a computer-readable storage medium, on which a computer program is stored. When the computer program is executed in a computer, the computer is made to execute the method described in combination with Figure 1 the above.

[0135] According to an embodiment of still another aspect, there is also provided a computing device, including a memory and a processor. An executable code is stored in the memory. When the processor executes the executable code, the method described in combination with Figure 1 the above is implemented.

[0136] Those skilled in the art should be able to realize that in one or more of the above examples, the functions described in the multiple embodiments disclosed in this specification can be implemented by hardware, software, firmware, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium.

[0137] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the multiple embodiments disclosed in this specification. It should be understood that the above are only the specific embodiments of the multiple embodiments disclosed in this specification, and are not used to limit the protection scope of the multiple embodiments disclosed in this specification. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solutions of the multiple embodiments disclosed in this specification shall be included within the protection scope of the multiple embodiments disclosed in this specification.

Claims

1. A method for constructing a model, wherein, the method includes: Receiving limited data input by a user based on an interaction interface for constructing a first model, where the first model is used to score each of a plurality of samples, and the receiving of the limited data for constructing the first model at least includes: Receiving at least one preset condition input, where each preset condition is a condition that the score result of the first model for at least one of the plurality of samples is required to satisfy; Obtaining at least one sample feature selected from a plurality of sample features for use as an input feature of the first model; Constructing the first model; An acquisition icon for the recommendation degree of sample features is further included in the interaction interface. After receiving the at least one preset condition input and before obtaining the at least one sample feature selected from the plurality of sample features, the method further includes: Receiving a trigger instruction for the acquisition icon; Displaying the recommendation degree of each of the plurality of sample features, where the recommendation degree is at least used to assist the user in selecting the at least one sample feature.

2. The method according to claim 1, wherein, Before receiving the limited data input by the user for constructing the first model, the method further includes: Receiving a selection instruction issued by the user based on the interaction interface for a first sample set, where the first sample set includes a plurality of samples; Displaying the sample data corresponding to each of the plurality of samples, where the sample data includes the feature values of each sample corresponding to each of the plurality of sample features.

3. The method according to claim 2, wherein, A sample set drop-down box is included in the interaction interface, and the sample set drop-down box includes a plurality of alternative sample sets, and the plurality of alternative sample sets include the first sample set; The receiving of the selection instruction issued by the user based on the interaction interface for the first sample set includes: Receiving the selection instruction issued by the user based on the sample set drop-down box for the first sample set.

4. The method according to claim 2, wherein, After displaying the sample data corresponding to each of the plurality of samples, the method further includes: In response to a folding instruction for the sample data, performing a folded display on the sample data.

5. The method according to claim 2, wherein, A normalization processing icon and a corresponding option box are included in the interaction interface. After displaying the sample data corresponding to each of the plurality of samples, the method further includes: In response to a confirmation selection instruction for the option box, updating and displaying the sample data by using the data obtained by performing normalization processing on the multiple feature values of the multiple samples corresponding to each of the sample features.

6. The method according to claim 1, wherein, The receiving of the at least one preset condition input includes: Receiving at least one preset condition set by the user.

7. The method according to claim 1, wherein, Among the at least one preset condition, there is a first preset condition whose condition type is the final classification type. The first preset condition includes the at least one sample, a first operator, and a preset score value. The first operator is one of greater than, less than, and equal to. The first preset condition is used to define the relative magnitude between the scores obtained by scoring each sample in the at least one sample using the first model and the preset score value.

8. The method according to claim 7, wherein, the interactive interface includes a condition type drop-down box, and the condition type drop-down box includes multiple alternative types, and the multiple alternative types include the final classification type; the receiving the input of the at least one preset condition includes: receiving a selection instruction for the final classification type issued based on the condition type drop-down box; displaying a first condition template corresponding to the final classification type. The first condition template includes a sample drop-down box, an operator drop-down box, and a calculated value input box. The sample drop-down box includes the multiple samples and is used to guide the user to select the at least one sample therefrom. The operator drop-down box includes multiple alternative operators and is used to guide the user to select the first operator therefrom. The calculated value input box is used to guide the user to input the preset score value; receiving a first predetermined condition input by the user based on the first condition template and incorporating it into the at least one preset condition.

9. The method according to claim 8, wherein, the first preset condition further includes a weight indicating the importance degree of this condition, and the first condition template further includes a weight input box for guiding the user to input the weight.

10. The method according to claim 1, wherein, among the at least one preset condition, there is a second preset condition whose condition type is the relative ranking type. The second preset condition includes a first sample subset, a second sample subset, and a second operator. The samples in the first sample subset and the second sample subset together constitute the at least one sample. The second operator is one of greater than, less than, and equal to. The second preset condition is used to define the relative magnitude between the scores obtained by scoring each sample in the first sample subset using the first model and the scores obtained by scoring each sample in the second sample subset.

11. The method according to claim 10, wherein, the interactive interface includes a condition type drop-down box, and the condition type drop-down box includes multiple alternative types, and the multiple alternative types include the relative ranking type; the receiving the input of the at least one preset condition includes: receiving a selection instruction for the relative ranking type issued based on the condition type drop-down box; Present a second conditional template corresponding to the relative ranking class, where the second conditional template includes a first sample dropdown box, a second sample dropdown box, and an operator dropdown box. The first sample dropdown box includes the multiple samples and is used to guide the user to select samples therefrom to form the first sample subset. The second sample dropdown box includes the multiple samples and is used to guide the user to select samples therefrom to form the second sample subset. The operator dropdown box includes multiple alternative operators and is used to guide the user to select the second operator therefrom; Receive the second predetermined condition input by the user based on the second conditional template and classify it into the at least one preset condition.

12. The method according to claim 11, wherein, The second preset condition further includes a weight indicating the importance degree of the condition, and the second conditional template further includes a weight input box for guiding the user to input the weight.

13. The method according to claim 1, wherein, The interactive interface includes multiple feature icons corresponding to the multiple sample features and filtering boxes corresponding to each of the feature icons. The obtaining of at least one sample feature selected from the multiple sample features includes: Receiving a confirmation selection instruction for at least one filtering box corresponding to the at least one sample feature.

14. The method according to claim 1, wherein, Receiving the limited data input by the user for constructing the first model further includes: Receiving at least one parameter value interval set for the at least one sample feature, and the at least one parameter value interval is used to limit the value range of the at least one model parameter.

15. The method according to claim 14, wherein, The interactive interface includes multiple adjustment icons for uniformly setting the upper and lower limits of the multiple parameter value intervals of the multiple sample features, including an upper limit uniformly increasing icon, an upper limit uniformly decreasing icon, a lower limit uniformly increasing icon, and a lower limit uniformly decreasing icon; Receiving at least one parameter value interval set for the at least one sample feature includes: In response to a trigger instruction for any first adjustment icon among the multiple adjustment icons, uniformly adjusting the multiple parameter value intervals corresponding to the first adjustment icon.

16. The method according to claim 1, wherein, The at least one sample feature includes a first sample feature, and the at least one model parameter includes a first model parameter corresponding to the first sample feature; Receiving the limited data input by the user for constructing the first model further includes: Receiving the first model parameter set for the first sample feature.

17. The method according to claim 16, wherein, The interactive interface includes multiple parameter value input boxes corresponding to the multiple sample features, including a first parameter value input box corresponding to the first sample feature; The receiving of the first model parameter set for the first sample feature includes: Receiving the first model parameter set based on the first parameter value input box.

18. The method according to claim 17, wherein, The interactive interface further includes a plurality of locking icons corresponding to the plurality of parameter value input boxes, including a first locking icon corresponding to the first parameter value input box, and the first locking icon is in an unlocked state; After receiving the first model parameter set based on the first parameter value input box, the method further includes: Receiving a trigger instruction for the first locking icon; Updating and displaying the first locking icon in the unlocked state as a locked state, and displaying the first parameter value input box displaying the first model parameter as non-editable.

19. The method according to claim 1, wherein, Receiving the limited data input by the user for constructing the first model further includes: Receiving the model type set for the first model, and the model type is a linear model or a non-linear model.

20. The method according to claim 1, wherein, After constructing the first model, the method further includes: Receiving a display instruction for the first model, and displaying the first model, where the first model includes the at least one sample feature and model parameters.

21. The method according to claim 1, wherein, The first model is a linear model, including at least one model parameter corresponding to the at least one sample feature, and the interactive interface includes at least one parameter value input box corresponding to the at least one sample feature. After constructing the first model, the method further includes: Displaying the at least one model parameter in the at least one parameter value input box.

22. The method according to claim 21, wherein, The at least one parameter value input box includes a first parameter value input box, and the at least one model parameter includes a first model parameter; After displaying the at least one model parameter in the at least one parameter value input box, the method further includes: Receiving the modification of the first model parameter by the user in the first parameter value input box for reconstructing the first model.

23. The method according to claim 1, wherein, After constructing the first model, the method further includes: Displaying a plurality of original scores obtained by scoring the plurality of samples using the first model.

24. The method according to claim 23, wherein, After constructing the first model, the method further includes: Displaying the original intervals corresponding to the plurality of original scores.

25. The method according to claim 24, wherein, After displaying the original intervals corresponding to the plurality of original scores, the method further includes: Receiving an adjusted interval set by the user based on the original interval; Replacing and displaying the plurality of original scores with a plurality of adjusted scores obtained by adjusting the plurality of original scores based on the adjusted interval.

26. The method according to claim 25, wherein, After receiving the adjusted interval set by the user based on the original interval, the method further includes: Displaying the gain value and offset value of the adjusted interval relative to the original interval.

27. The method according to claim 25, wherein, After building the first model and before receiving the adjusted interval set by the user based on the original interval, the method further includes: Displaying the original mean and original variance corresponding to the multiple original scores; After receiving the adjusted interval set by the user based on the original interval, the method further includes: Replacing and displaying the original mean and original variance with the adjusted mean and adjusted variance corresponding to the multiple adjusted scores, respectively.

28. The method according to claim 1, wherein, after building the first model, the method further includes: Displaying the multiple scores corresponding to the multiple samples, the multiple scores being determined based on the multiple original scores obtained by scoring the multiple samples using the first model.

29. The method according to claim 1, wherein, the interactive interface includes a histogram icon, and after building the first model, the method further includes: In response to a trigger instruction for the histogram icon, displaying a histogram drawn based on the multiple scores; wherein the histogram includes at least one histogram bar, and the numerical interval of each histogram bar relative to the abscissa represents the corresponding score interval, and the numerical value relative to the ordinate represents the corresponding number of samples.

30. The method according to claim 29, wherein, the histogram includes multiple reference grids, and the interval of each reference grid relative to the abscissa corresponds to a first interval score; after displaying the histogram drawn based on the multiple scores, the method further includes: Receiving a second interval score obtained by the user adjusting the first interval score; Updating and displaying the histogram using the second interval score.

31. The method according to claim 28, wherein, the displaying the multiple scores corresponding to the multiple samples includes: Displaying a ranking of the multiple samples based on the multiple scores; the interactive interface includes a radar chart icon, and after building the first model, the method further includes: Receiving a trigger instruction for the radar chart icon; Displaying at least one feature icon and at least one selection box corresponding to at least one sample feature, the at least one feature icon being arranged in order based on the magnitudes of the at least one model parameter, the at least one feature icon including a first number of selected feature icons, and the first number of selection boxes corresponding to the first number of feature icons being in a selected state; Displaying a radar chart drawn based on the first number of feature icons, the radar chart including a first number of radial axes; Receiving a second number of samples selected by the user from the multiple samples displayed in the ranking; Displaying a second number of closed curves corresponding to the second number of samples in the radar chart; Among them, the first quantity of radiation axes includes a first radiation axis corresponding to a first sample feature, the second quantity of closed curves includes a first closed curve corresponding to a first sample, and the value of the first closed curve on the first radiation axis is determined by the product of the weight value of the first sample feature and the feature value of the first sample corresponding to the first sample feature.

32. The method according to claim 1, wherein, the interactive interface further includes a preset condition verification icon, and after constructing the first model, the method further includes: responding to a trigger instruction for the preset condition verification icon, and presenting at least one condition difference corresponding to the at least one preset condition, wherein each condition difference reflects the compliance degree of the first model with respect to the preset condition corresponding to each condition difference.

33. An apparatus for constructing a model, wherein, the apparatus includes: a receiving unit configured to receive limited data input by a user based on an interactive interface for constructing a first model, the first model being used to score each of a plurality of samples, and the receiving unit at least includes: a first receiving module configured to receive at least one input preset condition, wherein each preset condition is a condition that requires the score result of the first model for at least one of the plurality of samples to meet; a second receiving module configured to obtain at least one sample feature selected from a plurality of sample features for use as an input feature of the first model; a constructing unit configured to construct the first model; the interactive interface further includes an acquisition icon for the recommendation degree of sample features, and the apparatus further includes: a third receiving mode configured to receive a trigger instruction for the acquisition icon; a presenting mode configured to present the recommendation degree of each sample feature among the plurality of sample features, and the recommendation degree is at least used to assist the user in selecting the at least one sample feature.

34. A computer-readable storage medium, on which a computer program is stored, wherein, when the computer program is executed on a computer, the computer is made to execute the method according to any one of claims 1-32.

35. A computing device, including a memory and a processor, wherein, an executable code is stored in the memory, and when the processor executes the executable code, the method according to any one of claims 1-32 is implemented.

Images