CN115659124A - Stability prediction method, apparatus, device and medium based on random sample weighting - Google Patents
Stability prediction method, apparatus, device and medium based on random sample weighting Download PDFInfo
- Publication number
- CN115659124A CN115659124A CN202211371406.1A CN202211371406A CN115659124A CN 115659124 A CN115659124 A CN 115659124A CN 202211371406 A CN202211371406 A CN 202211371406A CN 115659124 A CN115659124 A CN 115659124A
- Authority
- CN
- China
- Prior art keywords
- model
- prediction
- preset
- sample
- weighting
- 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
Abstract
The application relates to the technical field of machine learning, in particular to a stability prediction method, a stability prediction device, stability prediction equipment and stability prediction media based on random sample weighting, wherein the method comprises the following steps: randomly sampling model parameters in a preset function form from a preset distribution to obtain a plurality of weighted function models; inputting sample characteristics to each weighting function model, calculating sample weights corresponding to the weighting function models, and generating a plurality of heterogeneous environments based on the sample weights; and identifying a plurality of heterogeneous environments to obtain covariates of the prediction variables to obtain a stable prediction model reaching the preset condition. Therefore, the problems that sampling deviation in a training environment cannot be eliminated, stability and robustness of a model are influenced and the like in the related technology are solved.
Description
Technical Field
The present application relates to the field of machine learning technologies, and in particular, to a method, an apparatus, a device, and a medium for stable prediction based on random sample weighting.
Background
Machine learning models are being applied to high-risk fields such as automatic driving and intelligent medical treatment, and stability and robustness of the models are receiving more and more attention. When there is an unavoidable model misestimation, the stability and robustness of the model can be more challenging.
The traditional machine learning model is based on independent and identically distributed assumptions, and the experience risk loss of the model in a training sample is minimized. In a real scenario, when the test sample comes from a distribution different from the training sample, this assumption is not easily satisfied due to the unknown data source, and the covariate distribution has a deviation. Meanwhile, due to the assumed deviation of the model, the optimization of the model is more easily influenced by unstable relation and false correlation, and the systematic deviation can further amplify the selective deviation of the sample.
Disclosure of Invention
The application provides a method, a device, equipment and a medium for stable prediction based on random sample weighting, which are used for solving the problems that sampling deviation in a training environment cannot be eliminated, the stability and robustness of a model are influenced and the like in the related technology.
The embodiment of the first aspect of the present application provides a stable prediction method based on random sample weighting, including the following steps: randomly sampling model parameters in a preset function form from a preset distribution to obtain a plurality of weighted function models; inputting sample characteristics to each weighting function model, calculating sample weights corresponding to the weighting function models, and generating a plurality of heterogeneous environments based on the sample weights; and identifying the multiple heterogeneous environments to obtain covariates of the prediction variables to obtain a stable prediction model reaching preset conditions.
Optionally, in an embodiment of the present application, the generating a plurality of heterogeneous environments based on the sample weights includes: and under each heterogeneous environment, performing weighted calculation on all samples according to the sample weight obtained by each weighting function model to form a special case environment with distribution deviation, and generating the multiple heterogeneous environments.
Optionally, in an embodiment of the present application, the identifying covariates of the multiple heterogeneous environment acquisition predictor variables includes: and obtaining the prediction scores of the training samples and the stability indexes of the regression coefficients of the covariates in the multiple heterogeneous environments based on the objective function joint model to obtain the covariates.
Optionally, in an embodiment of the present application, the obtaining the stability indicator of the prediction score and the regression coefficient of the covariate of the training sample in the multiple heterogeneous environments includes: when the gradient optimization is carried out on the model each time, whether the objective function value of the model reaches a preset convergence condition is judged; and if the objective function value reaches the preset convergence condition, judging that the preset condition is reached.
The embodiment of the second aspect of the present application provides a stable prediction apparatus based on random sample weighting, including: the sampling module is used for randomly sampling model parameters in a preset function form from preset distribution to obtain a plurality of weighted function models; the generating module is used for inputting sample characteristics to each weighting function model, calculating sample weights corresponding to the weighting function models and generating a plurality of heterogeneous environments based on the sample weights; and the prediction module is used for identifying the multiple heterogeneous environments to obtain covariates of the prediction variables so as to obtain a stable prediction model reaching the preset condition.
Optionally, in an embodiment of the present application, the generating module is further configured to perform, in each heterogeneous environment, weighted calculation on all samples according to the sample weights obtained by each weighting function model to form a special case environment with distribution offset, and generate the multiple heterogeneous environments.
Optionally, in an embodiment of the present application, the prediction module includes: and the obtaining unit is used for obtaining the prediction scores of the training samples and the stability indexes of the regression coefficients of the covariates in the heterogeneous environments based on the target function joint model to obtain the covariates.
Optionally, in an embodiment of the present application, the obtaining unit is further configured to determine whether an objective function value of the model reaches a preset convergence condition each time the model performs gradient optimization; and if the objective function value reaches the preset convergence condition, judging that the preset condition is reached.
An embodiment of a third aspect of the present application provides an electronic device, including: a memory, a processor and a computer program stored on the memory and executable on the processor, the processor executing the program to implement the stable prediction method based on random sample weighting as described in the above embodiments.
A fourth aspect of the present application provides a computer-readable storage medium storing computer instructions for causing a computer to execute a stable prediction method based on random sample weighting according to the foregoing embodiment.
Therefore, the application has at least the following beneficial effects:
randomly sampling model parameters in a preset function form from a preset distribution to obtain a plurality of weighted function models; inputting sample characteristics to each weighting function model, calculating sample weights corresponding to the weighting function models, and generating a plurality of heterogeneous environments based on the sample weights; and identifying a plurality of heterogeneous environments to obtain covariates of the prediction variables to obtain a stable prediction model reaching the preset condition. The random sample weighting method based on the sample characteristics can generate countless random heterogeneous environments, identify stable covariates, relieve the amplification effect of the assumed deviation of the model on the sampling deviation and improve the prediction stability and robustness of the model. Therefore, the problems that sampling deviation in a training environment cannot be eliminated, stability and robustness of a model are influenced and the like in the related technology are solved.
Additional aspects and advantages of the present application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the present application.
Drawings
The above and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
fig. 1 is a flowchart of a stable prediction method based on random sample weighting according to an embodiment of the present application;
FIG. 2 is a schematic diagram of a stable prediction model provided according to an embodiment of the present application;
FIG. 3 is a diagram illustrating an example of a stable prediction apparatus based on random sample weighting according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Description of reference numerals: a sampling module-100, a generating module-200, a predicting module-300, a memory-401, a processor-402, and a communication interface-403.
Detailed Description
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary and intended to be used for explaining the present application and should not be construed as limiting the present application.
A method, an apparatus, a device, and a medium for stable prediction based on random sample weighting according to embodiments of the present application are described below with reference to the accompanying drawings. In view of the problems mentioned in the background art center, the present application provides a stable prediction method based on random sample weighting, in which a plurality of weighted function models are obtained by randomly sampling model parameters in a preset function form from a preset distribution; inputting sample characteristics to each weighting function model, calculating sample weights corresponding to the weighting function models, and generating a plurality of heterogeneous environments based on the sample weights; and identifying a plurality of heterogeneous environments to obtain covariates of the prediction variables to obtain a stable prediction model reaching the preset condition. Therefore, the problems that sampling deviation in a training environment cannot be eliminated, stability and robustness of a model are influenced and the like in the related technology are solved.
Specifically, fig. 1 is a schematic flowchart of a stable prediction based on random sample weighting according to an embodiment of the present disclosure.
As shown in fig. 1, the stable prediction method based on random sample weighting includes the following steps:
in step S101, model parameters in a preset function form are randomly sampled from a preset distribution to obtain a plurality of weighted function models.
Specifically, in order to calculate the weight of each sample feature, embodiments of the present application may randomly sample model parameters of a preset functional form (e.g., a generally common linear model) from a preset distribution (e.g., a generally common gaussian distribution or a uniform distribution), thereby obtaining a plurality of weighted function models. The preset function and distribution may be determined by those skilled in the art according to actual situations, and are not particularly limited.
In step S102, sample features are input to each weighting function model, sample weights corresponding to the weighting function models are calculated, and a plurality of heterogeneous environments are generated based on the sample weights.
Optionally, in an embodiment of the present application, generating a plurality of heterogeneous environments based on the sample weights includes: and under each heterogeneous environment, performing weighted calculation on all samples according to the sample weight obtained by each weighting function model to form a special case environment with distribution deviation, and generating a plurality of heterogeneous environments.
Specifically, the embodiment of the present application may calculate a sample weight by inputting a sample feature to each weighting function model (for example, when a linear model is used, an inner product of the sample feature and a model parameter is input as the sample weight), and one sample may correspond to different weights under different weighting functions. And under each environment, performing weighted calculation on all samples according to the corresponding weights obtained by the corresponding weighting functions to form special case environments with distribution offsets, thereby generating a plurality of heterogeneous environments. The embodiment of the application does not need a predefined heterogeneous environment, and the characteristic of each sample is subjected to calculation of a specific weight through each function. All samples under each function are subjected to specific weight calculation, and a special case environment with distribution deviation is formed through weighting, so that a plurality of heterogeneous environments are obtained.
It should be noted that, when the random heterogeneity environment is generated, the embodiment of the present application is not related to the back-end prediction model, and the bias influence of the model assumption on the sample weight learning is not introduced. Meanwhile, optimization and learning are not needed, and extra calculation cost is avoided.
In step S103, a plurality of heterogeneous environments are identified to obtain covariates of the predicted variables, and a stable prediction model meeting the preset conditions is obtained.
Optionally, in an embodiment of the present application, identifying a plurality of heterogeneous environment to obtain covariates of the predicted variable includes: and based on the target function joint model, obtaining the prediction scores of the training samples and the stability indexes of the regression coefficients of the covariates in a plurality of heterogeneous environments to obtain the covariates.
It can be understood that, in the embodiment of the present application, the prediction scores of the training samples and the stability indicators of the regression coefficients of the covariates in multiple heterogeneous environments (e.g., the update gradients of the regression coefficients of the variables in each environment) may be obtained, and the target covariate may be identified and obtained by determining which regression coefficients of the variables exhibit stability or invariance characteristics under the stability indicators (e.g., the regression coefficients of the variables are optimal solutions in multiple environments, and the update gradients are 0).
Specifically, as shown in fig. 2, the embodiment of the present application identifies covariates that maintain a stable relationship (have stable regression coefficients) to the predicted variables in multiple heterogeneous environments using an invariance learning technique: the objective function is combined with the stability index (such as gradient variance) of the prediction scores of the training samples and the regression coefficients of the covariates in a plurality of heterogeneous environments.
Optionally, in an embodiment of the present application, obtaining the stability indicator of the regression coefficients of the predictive score and the covariate of the training sample in multiple heterogeneous environments includes: when the gradient optimization is carried out on the model each time, whether the objective function value of the model reaches a preset convergence condition is judged; and if the target function value reaches the preset convergence condition, judging that the preset condition is reached.
It will be appreciated that each time the model is gradient optimized, all the above processes need to be performed cyclically until the objective function values converge, and a stable, robust predictive model can be obtained.
Taking a medical scenario as an example, the stable prediction method based on random sample weighting according to the present application will be described in detail.
In a medical setting, a model for detecting lung collapse in a chest x-ray film has been trained to demonstrate predictions based on the presence of chest tube drainage (a device commonly used during treatment). However, these models often experience prediction errors in cases without chest drainage. In the scenario of predicting the consumption of the electricity fee, the model easily considers irrelevant seasonal factors, such as the consumption of sun cream in summer, which causes performance differences in different seasons. The embodiment of the application amplifies fluctuation of unstable factors by self-generating a large number of heterogeneous environments, utilizes an invariance learning technology to identify stable covariates, improves prediction stability and robustness of the model, and obtains more stable performance effect in unknown test environments, such as not simply predicting lung collapse by using chest tube drainage and not considering seasonal factors to predict electricity charge.
According to the stability prediction method based on random sample weighting provided by the embodiment of the application, model parameters in a preset function form are randomly sampled from preset distribution to obtain a plurality of weighted function models; inputting sample characteristics to each weighting function model, calculating sample weights corresponding to the weighting function models, and generating a plurality of heterogeneous environments based on the sample weights; and identifying a plurality of heterogeneous environments to obtain covariates of the prediction variables to obtain a stable prediction model reaching the preset condition. The random sample weighting method based on the sample characteristics can generate countless random heterogeneous environments, identify stable covariates, relieve the amplification effect of the assumed deviation of the model on the sampling deviation and improve the prediction stability and robustness of the model. Therefore, the problems that sampling deviation in a training environment cannot be eliminated, stability and robustness of a model are influenced and the like in the related technology are solved.
Next, a stable prediction apparatus based on random sample weighting according to an embodiment of the present application will be described with reference to the drawings.
Fig. 3 is a block diagram illustrating a stable prediction apparatus based on random sample weighting according to an embodiment of the present application.
As shown in fig. 3, the stable prediction apparatus 10 based on random sample weighting includes: a sampling module 100, a generation module 200 and a prediction module 300.
The sampling module 100 is configured to randomly sample model parameters in a preset function form from a preset distribution to obtain a plurality of weighting function models. And a generating module 200, configured to input the sample characteristics to each weighting function model, calculate a sample weight corresponding to the weighting function model, and generate multiple heterogeneous environments based on the sample weight. The prediction module 300 is configured to identify covariates of multiple heterogeneous environments to obtain predicted variables, and obtain a stable prediction model meeting preset conditions.
Optionally, in an embodiment of the present application, the generating module 200 is further configured to, in each heterogeneous environment, perform a weighted calculation on all samples according to the sample weights obtained by each weighting function model to form a special case environment with distribution offset, so as to generate a plurality of heterogeneous environments.
Optionally, in an embodiment of the present application, the prediction module 300 includes: and the obtaining unit is used for obtaining the prediction scores of the training samples and the stability indexes of the regression coefficients of the covariates in a plurality of heterogeneous environments based on the target function combined model to obtain the covariates.
Optionally, in an embodiment of the present application, the obtaining unit is further configured to determine whether an objective function value of the model reaches a preset convergence condition each time the model performs gradient optimization; and if the objective function value reaches the preset convergence condition, judging that the objective function value reaches the preset condition.
It should be noted that the foregoing explanation of the embodiment of the stable prediction method based on random sample weighting is also applicable to the stable prediction apparatus based on random sample weighting in this embodiment, and is not repeated herein.
According to the stability prediction device based on random sample weighting provided by the embodiment of the application, model parameters in a preset function form are randomly sampled from preset distribution to obtain a plurality of weighted function models; inputting sample characteristics to each weighting function model, calculating sample weights corresponding to the weighting function models, and generating a plurality of heterogeneous environments based on the sample weights; and identifying a plurality of heterogeneous environments to obtain covariates of the prediction variables to obtain a stable prediction model reaching the preset condition. The random sample weighting method based on the sample characteristics can generate countless random heterogeneous environments, identify stable covariates, relieve the amplification effect of the assumed deviation of the model on the sampling deviation and improve the prediction stability and robustness of the model. Therefore, the problems that sampling deviation in a training environment cannot be eliminated, stability and robustness of a model are influenced and the like in the related technology are solved.
Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application. The electronic device may include:
The processor 402, when executing the program, implements the stable prediction method based on random sample weighting provided in the above embodiments.
Further, the electronic device further includes:
a communication interface 403 for communication between the memory 401 and the processor 402.
A memory 401 for storing computer programs operable on the processor 402.
If the memory 401, the processor 402 and the communication interface 403 are implemented independently, the communication interface 403, the memory 401 and the processor 402 may be connected to each other through a bus and perform communication with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended ISA (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown in FIG. 4, but this does not indicate only one bus or one type of bus.
Alternatively, in practical implementation, if the memory 401, the processor 402 and the communication interface 403 are integrated on a chip, the memory 401, the processor 402 and the communication interface 403 may complete communication with each other through an internal interface.
The processor 402 may be a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits configured to implement embodiments of the present Application.
The present embodiment also provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the above method of stable prediction based on random sample weighting.
In the description of the present specification, reference to the description of "one embodiment," "some embodiments," "an example," "a specific example," or "some examples" or the like means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or N embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present application, "N" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more N executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present application in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of implementing the embodiments of the present application.
The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or N wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present application may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc. Although embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application, and that variations, modifications, substitutions and alterations may be made to the above embodiments by those of ordinary skill in the art within the scope of the present application.
Claims (10)
1. A stable prediction method based on random sample weighting is characterized by comprising the following steps:
randomly sampling model parameters in a preset function form from a preset distribution to obtain a plurality of weighted function models;
inputting sample characteristics to each weighting function model, calculating sample weights corresponding to the weighting function models, and generating a plurality of heterogeneous environments based on the sample weights; and
and identifying the heterogeneous environments to obtain covariates of the prediction variables to obtain a stable prediction model reaching preset conditions.
2. The method as claimed in claim 1, wherein the generating a plurality of heterogeneous environments based on the sample weights comprises:
and under each heterogeneous environment, performing weighted calculation on all samples according to the sample weight obtained by each weighting function model to form a special case environment with distribution deviation, and generating the multiple heterogeneous environments.
3. The method as claimed in claim 1, wherein the identifying covariates of the plurality of heterogeneous environment acquisition predictor variables comprises:
and obtaining the prediction scores of the training samples and the stability indexes of the regression coefficients of the covariates in the multiple heterogeneous environments based on the objective function joint model to obtain the covariates.
4. The method as claimed in claim 3, wherein the obtaining of the stability indicator of the regression coefficients of the prediction scores and covariates of the training samples in the multiple heterogeneous environments comprises:
when the gradient optimization is carried out on the model each time, whether the objective function value of the model reaches a preset convergence condition is judged;
and if the objective function value reaches the preset convergence condition, judging that the preset condition is reached.
5. A stationary prediction apparatus based on random sample weighting, comprising:
the sampling module randomly samples model parameters in a preset function form from preset distribution to obtain a plurality of weighted function models;
the generating module is used for inputting sample characteristics to each weighting function model, calculating sample weights corresponding to the weighting function models and generating a plurality of heterogeneous environments based on the sample weights;
and the prediction module is used for identifying the multiple heterogeneous environments to obtain covariates of the prediction variables so as to obtain a stable prediction model reaching the preset condition.
6. The apparatus of claim 5, wherein the generating module is further configured to, in each heterogeneous environment, perform a weighted calculation on all samples according to the sample weights obtained by each weighting function model to form a special case environment with distributed offsets, and generate the plurality of heterogeneous environments.
7. The apparatus of claim 5, wherein the prediction module comprises:
and the obtaining unit is used for obtaining the prediction scores of the training samples and the stability indexes of the regression coefficients of the covariates in the heterogeneous environments based on the target function joint model to obtain the covariates.
8. The apparatus of claim 7, wherein the obtaining unit is further configured to determine whether an objective function value of the model reaches a preset convergence condition each time the model performs gradient optimization; and if the objective function value reaches the preset convergence condition, judging that the preset condition is reached.
9. An electronic device, comprising: memory, processor and computer program stored on the memory and executable on the processor, the processor executing the program to implement the method of stable prediction based on random sample weighting according to any of claims 1-4.
10. A computer-readable storage medium, on which a computer program is stored which is executable by a processor for implementing a method for stable prediction based on random sample weighting according to any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211371406.1A CN115659124A (en) | 2022-11-03 | 2022-11-03 | Stability prediction method, apparatus, device and medium based on random sample weighting |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211371406.1A CN115659124A (en) | 2022-11-03 | 2022-11-03 | Stability prediction method, apparatus, device and medium based on random sample weighting |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115659124A true CN115659124A (en) | 2023-01-31 |
Family
ID=84996308
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211371406.1A Pending CN115659124A (en) | 2022-11-03 | 2022-11-03 | Stability prediction method, apparatus, device and medium based on random sample weighting |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115659124A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117112445A (en) * | 2023-10-07 | 2023-11-24 | 太平金融科技服务（上海）有限公司 | Machine learning model stability detection method, device, equipment and medium |
-
2022
- 2022-11-03 CN CN202211371406.1A patent/CN115659124A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117112445A (en) * | 2023-10-07 | 2023-11-24 | 太平金融科技服务（上海）有限公司 | Machine learning model stability detection method, device, equipment and medium |
| CN117112445B (en) * | 2023-10-07 | 2024-01-16 | 太平金融科技服务（上海）有限公司 | Machine learning model stability detection method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Schubert et al. | Evaluating the model fit of diffusion models with the root mean square error of approximation | |
| CN111666993A (en) | Medical image sample screening method and device, computer equipment and storage medium | |
| GB2601929A (en) | A machine-learning based architecture search method for a neural network | |
| CN107402745B (en) | Mapping method and device of data flow graph | |
| CN115659124A (en) | Stability prediction method, apparatus, device and medium based on random sample weighting | |
| CN111914878B (en) | Feature point tracking training method and device, electronic equipment and storage medium | |
| CN110969200A (en) | Image target detection model training method and device based on consistency negative sample | |
| CN111639798A (en) | Intelligent prediction model selection method and device | |
| CN111859743A (en) | Structure dynamic displacement estimation method and device based on acceleration response | |
| CN114372412A (en) | Balanced tempering intelligent thawing method and system for low-temperature frozen food | |
| US20080255773A1 (en) | Machine condition monitoring using pattern rules | |
| CN111222553B (en) | Training data processing method and device of machine learning model and computer equipment | |
| CN114781248A (en) | Off-line reinforcement learning method and device based on state offset correction | |
| JP6764486B2 (en) | Water treatment plant | |
| CN111832693A (en) | Neural network layer operation and model training method, device and equipment | |
| CN114090952A (en) | Method, apparatus, device and storage medium for loss function dynamic weighting | |
| CN113671408A (en) | Leakage current detection method and detection device for lithium battery | |
| CN114399669B (en) | Target detection method and device | |
| CN117370796A (en) | Wind turbine generator system fault prediction model training and fault prediction method and device | |
| CN113988311B (en) | Quality variable prediction method, device, terminal and storage medium | |
| CN113127342B (en) | Defect prediction method and device based on power grid information system feature selection | |
| US20210326755A1 (en) | Learning model creation device, material property prediction device, and learning model creation method and program | |
| CN115879536A (en) | Learning cognition analysis model robustness optimization method based on causal effect | |
| KR20230135838A (en) | Method and apparatus for selective ensemble prediction based on dynamic model combination | |
| CN111047185B (en) | Method and device for determining influence of storage environment factors on ammunition storage reliability |
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 |